Tide-considered circulating water frequency conversion optimization control method for seawater direct cooling unit

Abstract

The invention provides a seawater direct cooling unit circulating water frequency conversion optimization control method considering tides, which provides a method for correcting a model of a frequency conversion circulating water pump by a seawater tide water level, establishes a circulating water pump optimization scheduling energy consumption model based on a micro output theory, and establishes a cold end optimization mathematical model for coupling a condenser and a steam turbine low-pressure cylinder; based on the Bayesian optimization algorithm, optimizing calculation of the target frequency of the optimal working point of parallel operation of the frequency conversion circulating pump is conducted, and the optimal operation mode of the circulating pump under the unit load, the circulating water inlet water temperature and the seawater level and a quick calculation method of the frequency conversion frequency are given; constructing a multiple regression model, and fitting a target frequency polynomial function by adopting a maximum likelihood method to simplify the workload of DCS logic configuration; the start-stop time of the circulating pump is reasonably arranged in combination with predicted data of the tidal water level, so that frequent start-stop of the circulating pump near a critical point and at a low water level is avoided, and equipment loss and energy consumption are reduced; and the seawater direct cooling unit circulating water frequency conversion optimization DCS closed-loop real-time optimization control is realized.

Description

Technical Field

[0001] The invention relates to the technical field of control of a seawater direct cooling unit, and in particular to a frequency conversion optimization control method for circulating water of a seawater direct cooling unit taking tide into consideration. Background Art

[0002] With a resource endowment dominated by coal and an installed capacity structure dominated by coal-fired power, thermal power is the basis for energy security, and it puts forward higher-level requirements for the operation optimization and energy conservation and consumption reduction of thermal power units.

[0003] The circulating water system is an important system in a thermal power plant. The operation of the circulating water system directly affects the condenser vacuum, which in turn affects the output of the steam turbine unit and the economy of the power plant. The circulating water pump is an important large-scale auxiliary machine in a thermal power plant. The electric energy consumed by the circulating water pump accounts for about 1% to 1.5% of the total power generation of the power plant and about 20% of the power consumption of the plant, which means there is a large space for energy saving. However, the management of the circulating water system in general power plants is relatively extensive, and it is difficult for operators to perform closed-loop real-time optimization and scheduling. Reasonable selection of the operating mode of the circulating water system to monitor its status in real time and optimize the scheduling of the circulating water pump is of great significance to improving the safety and economy of the power plant.

[0004] The circulating water pumps in thermal power plants have long been fixed-speed pumps. Since the fixed-speed circulation cannot directly adjust the circulating water flow, when the ambient temperature changes and the unit participates in peak load regulation, the circulating water volume often cannot match the actual load of the unit, causing the actual vacuum of the unit to deviate from the economic vacuum, which in turn affects the economic efficiency of the unit. In addition, since the characteristic curve of the fixed-speed pump cannot be changed, although the flow can be changed by throttling, this method will waste a lot of energy.

[0005] Under the background of rapid construction of new power systems, with the continuous improvement of new energy installed capacity and the power grid's ability to consume new energy electricity, thermal power units are facing increasing and deepening demands for deep peak regulation. At present, the flexibility transformation of thermal power is the key means to improve the regulation capacity of the power system and the most important source of incremental regulation capacity. At the same time, large-scale variable frequency speed regulation technology has been rapidly developed. In order to respond to the national energy policy requirements for coal-fired power units to save energy and reduce consumption, reduce plant power consumption, increase unit output, and further improve power plant economics, many power plants have begun to carry out variable frequency transformation of circulating water pumps, that is, to use variable frequency pumps instead of fixed speed pumps with high power consumption. Through variable frequency speed regulation transformation, the continuous adjustment of the circulating water pump speed and circulating water flow can be achieved, which provides a basis for further improving the optimization effect of the cold end system of the power plant. In addition, the operating power and plant power consumption rate of the auxiliary machine can be reduced, further reducing the impact of the auxiliary machine operation adjustment on the power grid and the unit. Therefore, studying how to finely adjust the circulating water pump operation combination mode and the inverter target frequency according to external factors, and then adjust the unit to the optimal vacuum operation, has become a key issue for energy saving and consumption reduction of the cold end system.

[0006] The seawater direct cooling thermal power unit is a direct cooling water system of a thermal power plant that uses seawater as a water source. With the continuous growth of energy demand and increasingly stringent environmental protection requirements, the application of seawater direct cooling units in coastal areas is becoming more and more extensive. However, due to the influence of tides, the seawater level will change periodically, which has a great impact on the circulating water system of the seawater direct cooling unit. Traditional circulating water control methods often fail to fully consider the changes in tides, resulting in low unit operation efficiency and increased energy consumption. For a long time, the circulating water management of seawater direct cooling units has been relatively extensive. When the circulating water system uses a fixed speed pump, most of the operating units of the circulating water pump are adjusted according to the season to adjust its operating status; when the circulating water system uses a variable frequency pump, most of the variable frequency frequencies are set according to the water temperature range with larger granularity and several typical loads. This cannot achieve the purpose of maintaining the optimal vacuum of the condenser. Due to the effect of ocean tides, seawater presents periodic fluctuations. Changes in seawater levels will affect the static pressure head of the circulating water pump pipeline characteristic curve, and then affect the pump head and the working point of the pump, and ultimately have a greater impact on the circulating water flow.

[0007] The change of tidal water level will directly affect the water level difference between the inlet and outlet of the circulating water system of the thermal power plant, thereby changing the head of the circulating pump. When the tidal water level rises, the head of the circulating pump decreases and the flow rate increases; conversely, when the tidal water level decreases, the head of the circulating pump increases and the flow rate decreases. Due to the instability of the circulating pump head and flow rate caused by the change of tidal water level, the circulating pump needs to constantly adjust its operating status to adapt to changes in external conditions. This not only increases the operating loss of the circulating pump, but also reduces its operating efficiency, thereby increasing the energy consumption of the circulating pump.

[0008] Therefore, for the open circulating water system of thermal power units directly cooled by seawater in coastal areas, the seawater tide level has a great influence on the circulating water flow and the power consumption of the water pump, and the two are constantly changing, with the daily high and low tide difference reaching 5 to 8 meters. Most of the existing related research contents on the research on the frequency conversion optimization of circulating water of seawater direct cooling units have ignored the important influence of tidal water level on the circulating water system, which will reduce the economy of the whole plant. The existing cold end optimization model established with unit load and circulating water inlet temperature as input factors is not suitable for the open circulating cold end system of seawater direct cooling units in coastal areas. Traditional circulating water pump optimization generally only uses the optimization results as recommended values ​​for operators to use, and does not incorporate the optimization strategy into the DCS or auxiliary control system to control the circulating water pump in real time. There is a limitation that the parameter optimization analysis results are out of touch with real-time control.

[0009] In addition, in the context of thermal power units widely participating in peak load regulation, the AGC load instructions issued by the dispatcher are frequently changing and fluctuating, and the seawater level will also change frequently and significantly within a day. Especially when the constraint condition is near the critical point, it is likely to cause the switching of multiple circulating pump combination modes, that is, the frequent start and stop of the circulating water pump and the frequent issuance of the target variable frequency instruction. Frequent start and stop of the circulating water pump takes a long time, which will cause large fluctuations in the vacuum and easily damage the equipment. Frequent changes in the operating speed of the variable frequency pump will also reduce the life of the equipment, which is not conducive to the economic operation of the unit.

[0010] In summary, facing the frequency conversion optimization and control problem of circulating water in seawater direct cooling units, relevant research has been conducted at home and abroad, but there are still some problems, such as the rapid calculation method for accurately determining the frequency conversion frequency, the lack of consideration of the influence of tidal water level on the optimization results, the lack of closed-loop real-time control of the optimization results, and the problem of frequent start and stop of water pumps and frequent changes in the operating speed of variable frequency pumps according to the optimal optimization results. To this end, a calculation method for frequency conversion optimization of circulating water in seawater direct cooling units considering tides is given, and the real-time optimization results are applied to the real-time control system of the unit to realize the real-time control of the optimal vacuum of the cold end system of the thermal power unit. This is of great significance to the operation optimization and energy saving and consumption reduction of thermal power units. Summary of the invention

[0011] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method for frequency conversion optimization control of circulating water in a seawater direct cooling unit taking tides into consideration. An energy consumption model for optimal scheduling of circulating water pumps is established based on the theory of slightly increased output, and a modeling method for the cold end optimization mathematical model of the coupling between the condenser and the low-pressure cylinder of the steam turbine is given. The optimal operation mode of the circulating pump under unit load, circulating water inlet water temperature, and seawater level and a method for rapid calculation of the frequency conversion frequency are given. After fitting the multiple optimization rules obtained under multiple constraints, a control strategy is written based on DCS, and measures are given to reduce the frequent start and stop of water pumps and the frequent changes in the operating speed of variable frequency pumps, ultimately achieving variable frequency closed-loop real-time control of the circulating water system, further improving the economy of thermal power plants.

[0012] In order to achieve the above object, the solution of the present invention is:

[0013] Step 100: Divide the circulating water pump combination mode according to the number of thermal power units, and divide the water level of the front water tank of the circulating water system into equal intervals according to the maximum and minimum seawater tide levels at the power plant location within a year. Preferably, the water level can be divided into intervals of 0.1m.

[0014] Step 200: Establish a mathematical model of the circulating pump when the variable frequency pump runs alone and in parallel with the fixed speed pump, divide according to the previous step to obtain the water level correction pipeline characteristic curve of the front water tank of the circulating water system, divide the circulating pump speed ratio into equal intervals, calculate the front water tank water level and the pump group working point under different combination modes, that is, the total flow rate of the circulating water pipe and the total power consumption of the circulating pump under a given circulating pump speed ratio, and use the least squares method to fit the functional relationship formula of the total flow rate of the circulating water pipe and the total power consumption of the circulating pump with the circulating pump speed ratio.

[0015] Step 300: Establish a cold end optimization mathematical model for the coupling of the condenser and the low-pressure cylinder of the turbine. Based on the theory of slightly increased output, the circulating water volume that maximizes the difference between the turbine's increased power and the circulating pump's power consumption is called the optimal circulating water volume. Determine the dividing line between the unit load and the circulating water inlet temperature under different circulating water pump combination modes at each front water tank water level.

[0016] Step 400: Taking the total power consumption of the pump group as the objective function of the circulating pump optimization energy consumption model, the target frequency of the variable frequency pump when different circulating water pump combination modes achieve the best vacuum at each front water tank water level is calculated based on the Bayesian optimization algorithm, and a multivariate regression model is constructed. The maximum likelihood method is used to fit the high-order polynomial function of the target frequency of the variable frequency pump and the unit load and the circulating water inlet water temperature.

[0017] Step 500: Based on the high-order polynomial function obtained by the above offline optimization, a control strategy algorithm is written in the DCS. The algorithm input is the water level of the front water tank, the unit load and the circulating water inlet water temperature. The output is the circulating water pump combination mode and the target frequency of the variable frequency pump. The maximum variable frequency range of the variable frequency circulating water pump is set. Preferably, the variable frequency range of the circulating water pump is 40Hz-50Hz. The target frequency of the variable frequency pump calculated by the DCS is smoothed and filtered, and the amplitude is limited and the change rate is limited. In addition, a certain dead zone and delay are set to reduce the frequent fluctuation and change of the frequency.

[0018] Step 600: Based on the ARIMA model, the water level of the front water tank, the unit load and the circulating water inlet temperature in the future specified time period are predicted. In order to reduce the frequent start and stop operations of the water pump that may occur when the unit is running near the critical point, the DCS control strategy is combined with the prediction model to perform pump cutting operations. Only when the input parameters calculated based on the cold end optimization mathematical model and the ARIMA model prediction show that there is no pump cutting operation in the future specified time period, and the DCS real-time control strategy determines that the water pump start-stop switching operation needs to be performed at the same time, the pump cutting operation is performed.

[0019] In step 100 of the above method:

[0020] The circulating water of thermal power units generally uses circulating water pumps of the same model, and the unit water supply system usually uses a combination of 1 unit and 2 pumps. Thermal power plants usually operate in the form of two or more units. This is to ensure that electricity can be continuously and stably supplied to users. When the demand of the power grid changes, the power grid needs can be more accurately met through operation modes such as startup, shutdown, and grid connection, thereby improving the stability and reliability of the power plant. For example, a thermal power plant has 2 units, and each unit circulating water system has 1 fixed speed pump and 1 variable frequency pump, and the two units have a total of 4 circulating water pumps of the same model. When the two thermal power units are in operation, the circulating water system has three conditions: 2 pumps, 3 pumps, and 4 pumps.

[0021] Specifically can be divided into

[0022] (1) Two fixed speed pumps running in parallel

[0023] (2) One fixed speed pump and one variable frequency pump run in parallel

[0024] (3) Two variable frequency pumps running in parallel

[0025] (4) Two variable speed pumps and one fixed speed pump run in parallel

[0026] (5) Two variable speed pumps and two fixed speed pumps running in parallel

[0027] In step 200 of the above method:

[0028] The variable frequency pump changes the actual speed of the pump, thereby changing the performance curve of the pump, and ultimately achieves the purpose of regulating the flow. When the variable frequency pump is running at the power frequency, the method for determining its working point is the same as that of the fixed speed circulating water pump. The calculation method of the fixed speed pump working point is as follows:

[0029] The working point of the constant speed circulating water pump is the intersection of the performance curve and the pipeline characteristic curve. The performance curve of the water pump is:

[0030] H=a 1 +a 2 Q+a 3 Q 2

[0031] In the formula, H is the pump head; Q is the pump flow rate; a 1 , a 2 , a 3 is a constant.

[0032] The pipeline characteristic curve of the water pump under the design head is:

[0033] H z =H s +γQ 2

[0034] In the formula, H z is the total resistance head, H s is the static pressure head, γ is a constant, and Q is the total flow rate in the pipeline.

[0035] The efficiency of the pump is:

[0036] η=b 1 Q+b 2 Q 2 +b 3 Q 3

[0037] Where b 1 , b 2 , b 3 is a constant.

[0038] The effective power of the pump is:

[0039] P=gHQ / η

[0040] In the formula, g is the gravitational acceleration constant.

[0041] When the frequency conversion is in operation, the characteristic curve of the water pump will change, and the intersection of the pipeline characteristic curve and the new performance curve is the new working point. The frequency conversion pump adjusts the speed of the pump by changing the frequency, thereby changing the performance curve of the pump to achieve flow regulation, and realizes continuous regulation of the circulating cooling water volume within the working range of the frequency converter.

[0042] The ratio of the actual speed of the variable frequency pump to the rated speed is called the speed ratio, as shown in the following formula:

[0043] ω=n i / n 0

[0044] Where n i is the actual speed of the pump, n 0 is the rated speed of the pump.

[0045] According to the similarity law, the actual head of the variable frequency pump is:

[0046] H=a 1 ω 2 +a 2 Q i +a 3 Q i 2

[0047] Where Q is the actual flow rate of the pump.

[0048] The operating efficiency of the pump can be obtained based on the actual flow rate of the variable frequency pump, and the actual power consumption of the variable frequency pump can be calculated based on the actual flow rate, actual head and efficiency.

[0049] When the sea water level changes, the system's pipeline characteristic curve will change.

[0050] H z =H s '+γQ 2

[0051] In the formula, H s ' is the corrected static pressure head, H s '=H s -ΔH, ΔH is the change in actual water level compared to the designed water level.

[0052] After the seawater level changes, the intersection of the new pipeline characteristic curve and the new performance curve is the new operating point of the pump.

[0053] When a fixed speed pump and a variable frequency pump are operated in parallel, different total flow rates will occur at different variable frequency pump speed ratios. When the variable frequency pump speed ratio is known, the calculation method for the system operating point is as follows: When two pumps are operated in parallel, the performance characteristic curve of the pump will change. A new performance characteristic curve is obtained by fitting the performance characteristic curve of the fixed speed pump with the performance characteristic curve of the variable frequency pump. The parallel performance curve is equivalent to the sum of the flow rates of the performance curves of each pump under the condition of equal head. The intersection of the new performance characteristic curve and the pipeline characteristic curve is the operating point. The actual flow rate and power consumption of each pump can be obtained based on the head of the operating point and the performance characteristic curves of each pump.

[0054] In step 300 of the above method:

[0055] In the energy consumption model, as the circulating water volume increases, the condenser vacuum rises, and the turbine increases its power output, but at the same time the power consumption of the circulating water pump increases accordingly. The benefit of low-compensation increased power is achieved, and the circulating water volume that maximizes the difference between the turbine increased power and the power consumption is called the optimal circulating water volume, and the corresponding vacuum is the optimal vacuum.

[0056] The objective function of the circulating pump optimization energy consumption model is:

[0057] max{ΔN=ΔN t -ΔN p}

[0058] Where: ΔN - the difference between the unit power increment and the circulating pump power increment, kW

[0059] ΔN t ——Unit power increment, kW

[0060] ΔN p——Increment of circulating pump power consumption, kW

[0061] Unit load change ΔN t

[0062] Determined based on the unit's general vacuum change and load change relationship curve

[0063]

[0064] The ratio of slightly increased power to output ΔN t / N t Only the ratio of back pressure to power p c / N t The relationship between them is called the universal curve of micro-increase power characteristic, which is applicable to different main steam flow rates or unit powers. According to the micro-increase power characteristic curves of steam turbines under different loads provided by the manufacturer, its universal curve can be further obtained.

[0065] In step 400 of the above method:

[0066] When the variable frequency pump is running, different speed ratios correspond to different flow rates and pump power consumption. When multiple pumps are running in parallel, for example, two variable frequency pumps are allowed to be connected in parallel. In theory, there are multiple combinations, that is, corresponding to multiple working points. Under each working point, the two pumps have their own speed ratios to meet the total flow requirements. However, there must be only one working point that minimizes the total power consumption of the pump group, that is, the objective function of the circulating pump optimization energy consumption model in step 300. The speed ratio corresponding to the working point at this time is the target speed ratio, and the corresponding frequency is the target frequency.

[0067] The optimal hyperparameters are found by constructing a probability model of the objective function. Bayesian optimization uses the Gaussian process as the probability model of the objective function and determines the next sampling point by maximizing the acquisition function. The calculation formula of the acquisition function is as follows:

[0068] a(x)=μ(x)+κσ(x)

[0069] Where a(x) is the value of the acquisition function, μ(x) is the Gaussian process model’s prediction of the mean of the target function at point x, σ(x) is the Gaussian process model’s prediction of the standard deviation of the target function at point x, and κ is a constant used to control the trade-off between exploration and exploitation.

[0070] The optimal circulation pump speed ratio and target frequency under the current working conditions are determined through the above Bayesian optimization.

[0071] In step 500 of the above method:

[0072] After the technical transformation of the circulating pump frequency conversion, if a single pump is running, it only needs to meet the vacuum requirements of the unit and the vibration requirements of the pump shaft system, and can run in the full frequency range. When the circulating water pump frequency conversion pump and the power frequency pump are running in parallel, the minimum frequency conversion frequency of the circulating water pump frequency conversion pump is affected by the vibration of the frequency conversion circulating water pump shaft system and the output limit of the frequency conversion pump when running in parallel with the power frequency pump, and according to the vibration requirements of the circulating water pump. After testing, and according to the vibration requirements of the circulating water pump, for safety reasons, the minimum frequency conversion range of the frequency conversion circulating water pump in the thermal logic is set to 40Hz, that is, the frequency conversion range of the circulating water pump is 40Hz-50Hz.

[0073] In step 600 of the above method:

[0074] Combined with the forecast data of tidal water level, the start and stop time of the circulation pump is reasonably arranged to avoid frequent start and stop of the circulation pump when the water level is low, reducing equipment loss and energy consumption. The ARIMA autoregressive moving average model is used to make short-term predictions of key system parameters, and the confidence interval prediction results at different confidence levels are combined with the deterministic prediction results to form a probabilistic prediction result.

[0075] Calculation using the formula

[0076]

[0077] where Y t is the predicted value, c is a constant, ε t and ε t-j is white noise, is the autoregressive coefficient, θ j is the sliding average coefficient, p is the order of the stationary time series autoregressive model, and q is the order of the moving average data model.

[0078] During offline optimization, the circulating water temperature division interval is at least 1°C. When the measured temperature is not at the divided temperature point, the interpolation method can be used to obtain the variable frequency pump target frequency corresponding to the measured temperature from the variable frequency pump target frequency corresponding to the adjacent temperature points.

[0079] The optimization results obtained by offline simulation modeling calculation need to be fitted to obtain the polynomial function corresponding to the three-dimensional surface of water level, unit load and circulating water inlet temperature to simplify the DCS configuration workload. The circulating water optimization control strategy runs periodically in the DCS, and the DCS and the inverter are connected by hard wiring to issue control instructions.

[0080] Beneficial effects: The present invention has the following advantages:

[0081] (1) Based on the theory of slightly increased output, an energy consumption model for optimal scheduling of circulating water pumps is established, and a modeling method for the cold end optimization mathematical model of the coupling between the condenser and the low-pressure cylinder of the steam turbine is given. The optimal operation mode of the circulating pump under unit load, circulating water inlet temperature, and seawater level and a fast calculation method for the frequency conversion frequency are given;

[0082] (2) Construct a multivariate regression model and use the maximum likelihood method to fit the high-order polynomial function of the variable frequency pump target frequency, unit load, and circulating water inlet temperature to supply the DCS to simplify the workload of DCS logic configuration;

[0083] (3) By accurately considering the impact of tides, the flow and pressure of circulating water are optimized. According to the calculation results of the optimization model, when the tidal water level changes, the speed, frequency and other operating parameters of the variable frequency circulating pump are adjusted in time to ensure that the circulating pump operates in the optimal state, thereby improving the power generation efficiency of the unit and effectively reducing energy consumption;

[0084] (4) Combined with the predicted data of tidal water levels, the start and stop time of the circulating pump is reasonably arranged, and a reasonable control strategy is designed to reduce the start and stop times and operating wear of equipment such as circulating water pumps, thereby extending the service life of the equipment; through optimized control, the unit can operate stably under complex conditions such as tidal changes, reducing equipment failures and downtime caused by operating condition fluctuations; ultimately, the DCS closed-loop real-time control of the circulating water frequency conversion optimization of the seawater direct cooling unit is realized, which significantly reduces the energy consumption of the variable frequency circulating pump of the thermal power plant, improves energy utilization efficiency and economic benefits, and plays a positive role in improving the energy utilization efficiency of thermal power plants, reducing operating costs and reducing environmental pollution. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 It is a flowchart of the algorithm execution of the present invention;

[0086] Figure 2 This is the relationship between the pump power and the circulating water flow rate and the speed ratio when the characteristic static pressure head of the pipeline is 10.5m for a single machine and a single pump;

[0087] Figure 3 This is the relationship between the pump power and circulating water flow rate and the speed ratio when the characteristic static pressure head of the pipeline is 6.5m for a single machine and a single pump;

[0088] Figure 4 This is the relationship between the pump power and circulating water flow rate and the speed ratio when the characteristic static pressure head of the pipeline under the dual-engine three-pump is 10.5m;

[0089] Figure 5 This is the relationship between the pump power and circulating water flow rate and the speed ratio when the characteristic static pressure head of the pipeline is 6.5m for two engines and three pumps;

[0090] Figure 6This is the relationship between the pump power and circulating water flow rate and the speed ratio when the characteristic static pressure head of the pipeline under the single machine and double pumps is 10.5m;

[0091] Figure 7 This is the relationship between the pump power and circulating water flow rate and the speed ratio when the characteristic static pressure head of the pipeline is 6.5m for a single machine with two pumps;

[0092] Figure 8 It is the dividing line of the circulating water pump combination mode when the pipeline characteristic static pressure head is 10.5m;

[0093] Fig. 9 It is the dividing line of the circulating water pump combination mode when the pipeline characteristic static pressure head is 6.5m;

[0094] Fig.10 It is the target frequency curve of the variable frequency pump at several typical temperatures in the single machine single pump mode when the pipeline characteristic static pressure head is 10.5m;

[0095] Fig.11 It is the fitting surface of the target frequency of the variable frequency pump in the single-unit single-pump mode and the unit load and circulating water inlet water temperature when the pipeline characteristic static pressure head is 10.5m;

[0096] Fig.12 It is the fitting surface of the target frequency of the variable frequency pump in the single-machine single-pump mode, the unit load and the circulating water inlet temperature when the pipeline characteristic static pressure head is 6.5m. DETAILED DESCRIPTION

[0097] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings.

[0098] like Figure 1 As shown, the method for frequency conversion optimization control of circulating water of a seawater direct cooling unit considering tides of the present invention comprises the following steps:

[0099] Step 100: Divide the circulating water pump combination mode according to the number of thermal power units, and divide the water level of the front water tank of the circulating water system into equal intervals according to the maximum and minimum seawater tide levels at the power plant location within a year. Preferably, the water level can be divided into intervals of 0.1m.

[0100] The circulating water of thermal power units generally uses circulating water pumps of the same model, and the unit water supply system usually uses a combination of 1 unit and 2 pumps. Thermal power plants usually operate in the form of two or more units to ensure that electricity can be continuously and stably supplied to users. When the demand of the power grid changes, the power grid needs can be more accurately met through operation modes such as startup, shutdown, and grid connection, thereby improving the stability and reliability of the power plant. For example, a thermal power plant has 2 units, and each unit unit circulating water system has 1 fixed speed pump and 1 variable frequency pump, and the two units have a total of 4 circulating water pumps of the same model. When the two thermal power units are in operation, the circulating water system has three conditions: 2 pumps, 3 pumps, and 4 pumps.

[0101] Specifically can be divided into

[0102] (1) Two fixed speed pumps running in parallel

[0103] (2) One fixed speed pump and one variable frequency pump run in parallel

[0104] (3) Two variable frequency pumps running in parallel

[0105] (4) Two variable speed pumps and one fixed speed pump run in parallel

[0106] (5) Two variable speed pumps and two fixed speed pumps operate in parallel.

[0107] Step 200: Establish a mathematical model of the circulating pump when the variable frequency pump runs alone and in parallel with the fixed speed pump, divide according to step 100 to obtain the water level correction pipeline characteristic curve of the front water tank of the circulating water system, divide the circulating pump speed ratio into equal intervals, calculate the front water tank water level and the pump group working point under different combination modes, that is, the total flow rate of the circulating water pipe and the total power consumption of the circulating pump under a given circulating pump speed ratio, and use the least squares method to fit the functional relationship between the total flow rate of the circulating water pipe and the total power consumption of the circulating pump and the circulating pump speed ratio.

[0108] The variable frequency pump changes the actual speed of the pump, thereby changing the performance curve of the pump, and ultimately achieves the purpose of regulating the flow. When the variable frequency pump is running at the power frequency, the method for determining its working point is the same as that of the fixed speed circulating water pump. The calculation method of the fixed speed pump working point is as follows:

[0109] The working point of the constant speed circulating water pump is the intersection of the performance curve and the pipeline characteristic curve. The performance curve of the water pump is:

[0110] H=a 1 +a 2 Q+a 3 Q 2

[0111] In the formula, H is the pump head; Q is the pump flow rate; a 1 , a 2 , a3 is a constant.

[0112] The pipeline characteristic curve of the water pump under the design head is:

[0113] H z =H s +γQ 2

[0114] In the formula, H z is the total resistance head, H s is the static pressure head, γ is a constant, and Q is the total flow rate in the pipeline.

[0115] The efficiency of the pump is:

[0116] η=b 1 Q+b 2 Q 2 +b 3 Q 3

[0117] Where b 1 , b 2 , b 3 is a constant.

[0118] The effective power of the pump is:

[0119] P=gHQ / η

[0120] In the formula, g is the gravitational acceleration constant.

[0121] When the frequency conversion is in operation, the characteristic curve of the water pump will change, and the intersection of the pipeline characteristic curve and the new performance curve is the new working point. The frequency conversion pump adjusts the speed of the pump by changing the frequency, thereby changing the performance curve of the pump to achieve flow regulation, and realizes continuous regulation of the circulating cooling water volume within the working range of the frequency converter.

[0122] The ratio of the actual speed of the variable frequency pump to the rated speed is called the speed ratio, as shown in the following formula:

[0123] ω=n i / n 0

[0124] Where n i is the actual speed of the pump, n 0 is the rated speed of the pump.

[0125] According to the similarity law, the actual head of the variable frequency pump is:

[0126] H=a 1 ω 2 +a 2 Q i +a3 Q i 2

[0127] Where Q is the actual flow rate of the pump.

[0128] The operating efficiency of the pump can be obtained based on the actual flow rate of the variable frequency pump, and the actual power consumption of the variable frequency pump can be calculated based on the actual flow rate, actual head and efficiency.

[0129] When the sea water level changes, the system's pipeline characteristic curve will change.

[0130] H z =H s '+γQ 2

[0131] In the formula, H s ' is the corrected static pressure head, H s '=H s -ΔH, ΔH is the change in actual water level compared to the designed water level.

[0132] After the seawater level changes, the intersection of the new pipeline characteristic curve and the new performance curve is the new operating point of the pump.

[0133] When a fixed speed pump and a variable frequency pump are operated in parallel, different total flow rates will occur at different variable frequency pump speed ratios. When the variable frequency pump speed ratio is known, the calculation method for the system operating point is as follows: When two pumps are operated in parallel, the performance characteristic curve of the pump will change. A new performance characteristic curve is obtained by fitting the performance characteristic curve of the fixed speed pump with the performance characteristic curve of the variable frequency pump. The parallel performance curve is equivalent to the sum of the flow rates of the performance curves of each pump under the condition of equal head. The intersection of the new performance characteristic curve and the pipeline characteristic curve is the operating point. The actual flow rate and power consumption of each pump can be obtained based on the head of the operating point and the performance characteristic curves of each pump.

[0134] Step 300: Establish an optimized energy consumption model for the circulating pump coupled with the condenser and the low-pressure cylinder of the turbine. Based on the theory of slightly increased output, the circulating water volume that maximizes the difference between the turbine's increased power and the circulating pump's power consumption is called the optimal circulating water volume. Determine the dividing line between the unit load and the circulating water inlet temperature under different circulating water pump combination modes at each front water tank water level.

[0135] In the energy consumption model, as the circulating water volume increases, the condenser vacuum rises, and the turbine increases its power output, but at the same time the power consumption of the circulating water pump increases accordingly. The benefit of low-compensation increased power is achieved, and the circulating water volume that maximizes the difference between the turbine increased power and the power consumption is called the optimal circulating water volume, and the corresponding vacuum is the optimal vacuum.

[0136] The objective function of the circulating pump optimization energy consumption model is:

[0137] max{ΔN=ΔN t -ΔN p}

[0138] Where: ΔN - the difference between the unit power increment and the circulating pump power increment, kW

[0139] ΔN t ——Unit power increment, kW

[0140] ΔN p ——Increment of circulating pump power consumption, kW

[0141] 1) Condensate flow rate D c

[0142] Ignore the auxiliary steam entering the condenser, condensation volume D c is the exhaust volume of low pressure cylinder G s With the small steam engine exhaust volume G st According to the exhaust volume of low pressure cylinder of #1 engine G s With the small steam engine exhaust volume G st The sum of the design parameters under no load is fitted to obtain the condensation volume D c With unit load N t Mathematical relationship of (MW):

[0143] D c =1.481508N t +125.3772

[0144] 2) Circulating water temperature rise Δt

[0145] For single pressure condenser

[0146]

[0147] Where: D c —Condenser steam condensation capacity, t / h

[0148] Δh—the difference between the specific enthalpy of steam and the specific enthalpy of condensed water in the condenser, kJ / kg, the initial value is generally 2200;

[0149] D w —Cooling water flow rate, t / h, is determined according to the operating status of the circulating pump.

[0150] When a power plant uses a double-pressure condenser, the cooling water temperature rises at the cold and hot ends of the double-pressure condenser.

[0151]

[0152] 3) Exhaust steam temperature t c and exhaust steam pressure P c

[0153] For a single-pressure condenser, the exhaust temperature t c The following formula is used for calculation:

[0154] t c =t 1 +Δt+δt

[0155] Where: t 1 —Circulating water inlet temperature, °C; Δt—circulating water temperature rise, t 2 -t 1 , ℃; δt—condenser end heat transfer difference, ℃.

[0156]

[0157] Where: K is the overall heat transfer coefficient of the condenser, kJ / (㎡·h·℃); A c —Total outer surface area of ​​cooling water pipe, m2.

[0158] The heat transfer difference between the cold and hot ends of the double-pressure condenser is

[0159]

[0160] The condenser heat transfer coefficient K is calculated using the Berman formula:

[0161] K=14650φφ w φ t φ z φ d

[0162]

[0163]

[0164]

[0165] Where φ is the correction factor of cooling surface cleanliness, i.e., cleanliness factor. In this paper, φ = 0.85;

[0166] φ w - Correction factors for cooling water velocity and pipe diameter;

[0167] c w ——The flow rate of cooling water in the pipe;

[0168] d 1 ——Inner diameter of cooling water pipe;

[0169] φ t ——Cooling water inlet temperature correction coefficient;

[0170] φ z ——Correction coefficient of cooling water flow number Z;

[0171] φ d ——Steam load per unit area of ​​condenser d c The amount of steam condensed per unit area per unit time is called unit heat load, i.e. d c =D c / A c When d c In design value With critical value

[0172] Between When no correction is required, φ d =1.

[0173] when When d =δ(2-δ),

[0174] The condensation temperatures at the cold and hot ends of the double-pressure condenser are

[0175]

[0176] t s2 =t w1 +Δt+δt 2

[0177] Average condensing temperature of double pressure condenser

[0178]

[0179] Depend on The corresponding saturated steam pressure, i.e., the condenser pressure p, can be determined. c .

[0180] 4) Unit load change ΔN t

[0181] Determined based on the unit's general vacuum change and load change relationship curve

[0182]

[0183] The ratio of slightly increased power to output ΔN t / N t Only the ratio of back pressure to power p c / N t The relationship between them is called the universal curve of micro-increase power characteristic, which is applicable to different main steam flow rates or unit powers. According to the micro-increase power characteristic curve under different loads of the steam turbine provided by the manufacturer, its universal curve can be further obtained. t / N t With p c / N t It is a linear relationship, and the following expression is obtained after fitting:

[0184]

[0185] In the formula, p c ——Back pressure (condenser pressure), kPa; N t ——Unit power, MW.

[0186] Step 400: Taking the total power consumption of the pump group as the objective function of the circulating pump optimization energy consumption model, the target frequency of the variable frequency pump when different circulating water pump combination modes achieve the best vacuum at each front water tank water level is calculated based on the Bayesian optimization algorithm, and a multivariate regression model is constructed. The maximum likelihood method is used to fit the high-order polynomial function of the target frequency of the variable frequency pump and the unit load and the circulating water inlet water temperature.

[0187] When the variable frequency pump is running, different speed ratios correspond to different flow rates and pump power consumption. When multiple pumps are running in parallel, for example, two variable frequency pumps are running in parallel, there are theoretically multiple combinations, that is, corresponding to multiple working points, and the two pumps have their own speed ratios at each working point to meet the total flow requirements. However, there must be only one working point that minimizes the total power consumption of the pump group, which is the objective function of the circulating pump optimization energy consumption model in step 300. The speed ratio corresponding to the working point at this time is the target speed ratio, and the corresponding frequency is the target frequency.

[0188] The optimal hyperparameters are found by constructing a probability model of the objective function. Bayesian optimization uses the Gaussian process as the probability model of the objective function and determines the next sampling point by maximizing the acquisition function. The calculation formula of the acquisition function is as follows:

[0189] a(x)=μ(x)+κσ(x)

[0190] Where a(x) is the value of the acquisition function, μ(x) is the Gaussian process model’s prediction of the mean of the target function at point x, σ(x) is the Gaussian process model’s prediction of the standard deviation of the target function at point x, and κ is a constant used to control the trade-off between exploration and exploitation.

[0191] The optimal circulation pump speed ratio and target frequency under the current working conditions are determined through the above Bayesian optimization.

[0192] Step 500: Based on the high-order polynomial function obtained by the above offline optimization, a control strategy algorithm is written in the DCS. The algorithm input is the water level of the front water tank, the unit load and the circulating water inlet water temperature. The output is the circulating water pump combination mode and the target frequency of the variable frequency pump. The maximum variable frequency range of the variable frequency circulating water pump is set. Preferably, the variable frequency range of the circulating water pump is 40Hz-50Hz. The target frequency of the variable frequency pump calculated by the DCS is smoothed and filtered, and the amplitude is limited and the change rate is limited. In addition, a certain dead zone and delay are set to reduce the frequent fluctuation and change of the frequency.

[0193] During offline optimization, the circulating water temperature division interval is at least 1°C. In this embodiment, the circulating water temperature division interval during offline optimization is 1°C. When the measured temperature is not at the division temperature point, the interpolation method can be used to obtain the variable frequency pump target frequency corresponding to the measured temperature from the variable frequency pump target frequency corresponding to the adjacent temperature points.

[0194] After the technical transformation of the circulating pump frequency conversion, if a single pump is running, it only needs to meet the vacuum requirements of the unit and the vibration requirements of the pump shaft system, and can run in the full frequency range. When the circulating water pump and the power frequency pump are running in parallel, the minimum frequency conversion frequency of the circulating water pump is affected by the vibration of the variable frequency circulating water pump shaft system and the output limit of the variable frequency pump when running in parallel with the power frequency pump, and meets the vibration requirements of the circulating water pump. For safety reasons, the minimum frequency conversion range of the variable frequency circulating water pump in the thermal logic is set to 40Hz, that is, the frequency conversion range of the circulating water pump is 40Hz-50Hz.

[0195] Step 600: Based on the ARIMA model, the water level of the front water tank, the unit load and the circulating water inlet temperature in the future specified time period are predicted. In order to reduce the frequent start and stop operations of the water pump that may occur when the unit is running near the critical point, the DCS control strategy is combined with the prediction model to perform pump cutting operations. Only when the input parameters calculated based on the cold end optimization mathematical model and the ARIMA model prediction show that there is no pump cutting operation in the future specified time period, and the DCS real-time control strategy determines that the water pump start-stop switching operation needs to be performed at the same time, the pump cutting operation is performed.

[0196] Combined with the forecast data of tidal water level, the start and stop time of the circulation pump is reasonably arranged to avoid frequent start and stop of the circulation pump when the water level is low, reducing equipment loss and energy consumption. The ARIMA autoregressive moving average model is used to make short-term forecasts of key system parameters, and the confidence interval forecast results at different confidence levels are given and combined with the deterministic forecast results to form a probabilistic forecast result. The following formula is used for calculation:

[0197]

[0198] where Y t is the predicted value, c is a constant, ε t and εt-j is white noise, is the autoregressive coefficient, θ j is the sliding average coefficient, p is the order of the stationary time series autoregressive model, and q is the order of the moving average data model.

[0199] The present invention establishes an energy consumption model for optimizing the scheduling of circulating water pumps based on the theory of micro-increase output, provides a modeling method for the cold-end optimization mathematical model of the coupling between the condenser and the low-pressure cylinder of the steam turbine, and provides the optimal operation mode of the circulating pump under the unit load, circulating water inlet water temperature, and seawater level, as well as a method for quickly calculating the variable frequency. After fitting the multiple optimization rules obtained under the multi-constraint conditions, a control strategy is written based on DCS, and measures are provided to reduce the frequent start and stop of the water pump and the frequent change of the operating speed of the variable frequency pump, ultimately realizing the variable frequency closed-loop real-time control of the circulating water system, further improving the economy of the thermal power plant.

[0200] Example

[0201] This embodiment establishes a 2*660MW thermal power unit variable frequency speed regulation circulating water pump modeling and optimization control research model to optimize the control of relevant parameters. Two 660MW units in a power plant are taken as the research objects. The steam turbine of a power plant is a 660MW ultra-supercritical, single intermediate reheat, single shaft, four-cylinder four-exhaust, double back pressure, condensing steam turbine. The type of condenser is double back pressure, double shell, single flow, surface type, horizontal arrangement, which consists of a condenser on the high pressure side and a condenser on the low pressure side. This condenser is designed according to the TMCR working condition of the steam turbine. When the cooling water temperature is 20℃, the average back pressure of the condenser is 4.9kPa. The oxygen content of the condensate at the condenser outlet does not exceed 20ppb during normal operating conditions. The circulating cooling water enters the condenser from the inlet of the low pressure side inlet and outlet water chamber in two routes, and then from the return water chamber to the high pressure side shell. After heat exchange with steam in the two shells, it flows out from the inlet and outlet water chambers on the high pressure side. The circulating water adopts an expanded unit direct current water supply system to provide cooling water for the condenser and open circulating cooling water system. Each unit is equipped with two vertical single-stage, rotor-removable guide vane mixed flow pumps with a flow rate of Q = 10.5m3 / s and a head of H = 18m; the matching motor power is 2750kW. One circulating water pump of each unit is a variable frequency pump, and the other circulating water pump is a fixed speed pump.

[0202] like Figure 1 As shown, the process of this embodiment includes the following steps:

[0203] Step 100: Divide the circulating water pump combination mode according to the number of thermal power units, and divide the water level of the front water tank of the circulating water system into equal intervals according to the maximum and minimum seawater tide levels at the power plant location within a year. Preferably, the water level can be divided into intervals of 0.1m.

[0204] Considering only the case where the communication door of the circulating water pump room is open, when both 660MW units are in operation, there are three circulating pump combination modes: dual-machine dual-pump (one variable frequency circulating pump for each unit), dual-machine three-pump (one fixed speed pump + half variable frequency circulating pump for each unit) and dual-machine four-pump (one fixed speed pump + one variable frequency circulating pump for each unit). When only one 660MW unit is in operation, there are two circulating pump combination modes: single-machine single-pump (one variable frequency circulating pump) and single-machine dual-pump (one fixed speed pump + one variable frequency circulating pump). Among them, the optimization effect of dual-machine dual-pump is the same as that of single-machine single-pump, and the optimization effect of dual-machine four-pump is the same as that of single-machine dual-pump.

[0205] The input of the control model is the current unit load / MW, circulating water inlet temperature / ℃, and front water tank water level / m. The output of the model is the circulating pump combination mode (one of three modes), that is, the target frequency of the variable frequency pump (37.5-50Hz).

[0206] The water level of the front water tank changes in the range of 7-12 meters in a year. In this calculation process, the range of the front water tank is 6-13 meters, and the water level is divided into 8 parts for cold end optimization calculation according to an interval of 1 meter. The water levels are 6, 7, 8, 9, 10, 11, 12, and 13 meters respectively.

[0207] Step 200: Establish a mathematical model of the circulating pump when the variable frequency pump runs alone and in parallel with the fixed speed pump, divide according to the previous step to obtain the water level correction pipeline characteristic curve of the front water tank of the circulating water system, divide the circulating pump speed ratio into equal intervals, calculate the front water tank water level and the pump group working point under different combination modes, that is, the total flow rate of the circulating water pipe and the total power consumption of the circulating pump under a given circulating pump speed ratio, and use the least squares method to fit the functional relationship formula of the total flow rate of the circulating water pipe and the total power consumption of the circulating pump with the circulating pump speed ratio.

[0208] The flow meter power consumption characteristic data of the fixed speed pump and the variable speed pump in the dual-machine three-pump mode when the pipeline characteristic static pressure head is 11.5m are shown in Table 1.

[0209] Table 1 Calculation data of water pump characteristics for dual-engine three-pump system when pipeline characteristic static pressure head is 11.5m

[0210]

[0211]

[0212] The relationship between the pump power, circulating water flow rate and speed ratio when the pipeline characteristic static pressure head is 10.5m and 6.5m in the three circulating water pump combination modes of single machine single pump, dual machine three pumps and single machine two pumps is as follows Figures 2 to 7 shown.

[0213] Step 300: Establish a cold end optimization mathematical model for the coupling of the condenser and the low-pressure cylinder of the turbine. Based on the theory of slightly increased output, the circulating water volume that maximizes the difference between the turbine's increased power and the circulating pump's power consumption is called the optimal circulating water volume. Determine the dividing line between the unit load and the circulating water inlet temperature under different circulating water pump combination modes at each front water tank water level.

[0214] Since the surface seawater temperature is

[0215] In winter, the average temperature of each sea area is: Yellow Sea 6.72℃,

[0216] In summer, the average temperature in various sea areas is: 25.67℃ in the Yellow Sea.

[0217] In this example, the inlet circulating water temperature range is 4 to 35°C.

[0218] The pipeline characteristic curve is H z =H s +Q 2

[0219] The change of water level in the front water tank only affects the static pressure head of the pipeline characteristic curve, and does not affect the performance characteristic curve.

[0220] The height of the front water tank is 15 meters, the elevation of the condenser circulating water inlet is 2.5 meters, and the water level of the front water tank is H i , then the static pressure head is H s =17.5-H i .

[0221] The water level of the front water tank changes in the range of 8-12 meters in a year. In this calculation process, the range of the front water tank is 6-13 meters, and the water level is divided into 8 parts for cold end optimization calculation according to an interval of 1 meter. The water levels are 6, 7, 8, 9, 10, 11, 12, and 13 meters respectively.

[0222] The pipeline characteristic static pressure head is 11.5m, corresponding to the front water tank water level of 6m

[0223] The pipeline characteristic static pressure head is 10.5m, corresponding to the front water tank water level of 7m

[0224] The pipeline characteristic static pressure head is 9.5m, corresponding to the front water tank water level of 8m

[0225] The pipeline characteristic static pressure head is 8.5m, corresponding to the front water tank water level of 9m

[0226] The pipeline characteristic static pressure head is 7.5m, corresponding to the front water tank water level of 10m

[0227] The pipeline characteristic static pressure head is 6.5m, corresponding to the front water tank water level of 11m

[0228] The pipeline characteristic static pressure head is 5.5m, corresponding to the front water tank water level of 12m

[0229] The pipeline characteristic static pressure head is 4.5m, corresponding to the front water tank water level of 13m

[0230] The dividing line between a single machine and a single pump and a double machine and three pumps is shown below. There are 8 curves in total, and the corresponding static pressure head of the pipeline characteristic curve is △Z = 11.5 ~ 4.5, with an interval of 1 meter. The water levels of the front water tanks corresponding to the 8 curves are 6, 7, 8, 9, 10, 11, 12, and 13 meters respectively.

[0231] y=6E-05*t*t-0.1156*t+67.606;_

[0232] y=6E-05*t*t-0.1129*t+66.626;

[0233] y=6E-05*t*t-0.1077*t+64.867;

[0234] y=6E-05*t*t-0.1067*t+64.31;_

[0235] y=6E-05*t*t-0.1052*t+63.606;_

[0236] y=5E-05*t*t-0.1032*t+62.802;_

[0237] y=5E-05*t*t-0.1014*t+61.843;_

[0238] y=5E-05*t*t-0.1005*t+61.334;_

[0239] Note: t represents load / MW, y represents inlet circulating water temperature / ℃. For example,

[0240] When the water level is less than 6 meters, the curve y=6E-05*t*t-0.1156*t+67.606 is used;

[0241] When the water level is greater than 13 meters, the curve y=5E-05*t*t-0.1005*t+61.334 is used;

[0242] The dividing line between the dual-machine three-pump and single-machine two-pump is shown below. There are 8 curves in total, and the corresponding static pressure head of the pipeline characteristic curve is △Z = 11.5 ~ 4.5, with an interval of 1 meter. The water levels of the front water tanks corresponding to the 8 curves are 6, 7, 8, 9, 10, 11, 12, and 13 meters respectively.

[0243] y=8E-05*t*t-0.146*t+83.122;_

[0244] y=8E-05*t*t-0.1433*t+83.216;_

[0245] y=9E-05*t*t-0.1493*t+85.447;_

[0246] y=9E-05*t*t-0.1529*t+86.822;_

[0247] y=8E-05*t*t-0.1449*t+84.627;_

[0248] y=8E-05*t*t-0.1455*t+84.993;_

[0249] y=9E-05*t*t-0.1533*t+86.188;_

[0250] y=8E-05*t*t-0.1484*t+87.646;

[0251] The dividing line between the circulating water pump combination mode when the pipeline characteristic static pressure head is 10.5m and 6.5m is as follows Figures 8-9 shown.

[0252] It can be seen that when the unit load and the circulating water inlet temperature are both high, the single-machine dual-pump mode is adopted; when the unit load and the circulating water inlet temperature are both low, the single-machine single-pump mode is adopted; in other cases, the dual-machine three-pump mode is adopted.

[0253] Step 400: Taking the total power consumption of the pump group as the objective function of the circulating pump optimization energy consumption model, the target frequency of the variable frequency pump when different circulating water pump combination modes achieve the best vacuum at each front water tank water level is calculated based on the Bayesian optimization algorithm, and a multivariate regression model is constructed. The maximum likelihood method is used to fit the high-order polynomial function of the target frequency of the variable frequency pump and the unit load and the circulating water inlet water temperature.

[0254] The optimal hyperparameters are found by constructing a probability model of the objective function. Bayesian optimization uses the Gaussian process as the probability model of the objective function and determines the next sampling point by maximizing the acquisition function. The calculation formula of the acquisition function is as follows:

[0255] a(x)=μ(x)+κσ(x)

[0256] Where a(x) is the value of the acquisition function, μ(x) is the Gaussian process model’s prediction of the mean of the target function at point x, σ(x) is the Gaussian process model’s prediction of the standard deviation of the target function at point x, and κ is a constant used to control the trade-off between exploration and exploitation.

[0257] When the pipeline characteristic static pressure head is 10.5m, the target frequency curve of the variable frequency pump at several typical temperatures in the single machine single pump mode is as follows Fig.10 As shown; the fitting surface of the target frequency of the variable frequency pump in the single-machine single-pump mode and the unit load and circulating water inlet temperature when the pipeline characteristic static pressure head is 10.5m and 6.5m is as follows Figures 11-12 shown.

[0258] Step 500: Based on the high-order polynomial function obtained by the above offline optimization, a control strategy algorithm is written in the DCS. The algorithm input is the water level of the front water tank, the unit load and the circulating water inlet water temperature. The output is the circulating water pump combination mode and the target frequency of the variable frequency pump. The maximum variable frequency range of the variable frequency circulating water pump is set. Preferably, the variable frequency range of the circulating water pump is 40Hz-50Hz. The target frequency of the variable frequency pump calculated by the DCS is smoothed and filtered, and the amplitude is limited and the change rate is limited. In addition, a certain dead zone and delay are set to reduce the frequent fluctuation and change of the frequency.

[0259] After the technical transformation of the circulating pump frequency conversion, if a single pump is running, it only needs to meet the vacuum requirements of the unit and the vibration requirements of the pump shaft system, and can run in the full frequency range. When the circulating water pump frequency conversion pump and the power frequency pump are running in parallel, the minimum frequency conversion frequency of the circulating water pump frequency conversion pump is affected by the vibration of the frequency conversion circulating water pump shaft system and the output limit of the frequency conversion pump when running in parallel with the power frequency pump, and according to the vibration requirements of the circulating water pump. After testing, and according to the vibration requirements of the circulating water pump, for safety reasons, the minimum frequency conversion range of the frequency conversion circulating water pump in the thermal logic is set to 40Hz, that is, the frequency conversion range of the circulating water pump is 40Hz-50Hz.

[0260] Step 600: Based on the ARIMA model, the water level of the front water tank, the unit load and the circulating water inlet temperature in the future specified time period are predicted. In order to reduce the frequent start and stop operations of the water pump that may occur when the unit is running near the critical point, the DCS control strategy is combined with the prediction model to perform pump cutting operations. Only when the input parameters calculated based on the cold end optimization mathematical model and the ARIMA model prediction show that there is no pump cutting operation in the future specified time period, and the DCS real-time control strategy determines that the water pump start-stop switching operation needs to be performed at the same time, the pump cutting operation is performed.

[0261] The ARIMA autoregressive moving average model is used to make short-term forecasts of key system parameters, and the confidence interval forecast results at different confidence levels are combined with the deterministic forecast results to form a probabilistic forecast result.

[0262] Calculation using the formula

[0263]

[0264] where Y t is the predicted value, c is a constant, εt and ε t-j is white noise, is the autoregressive coefficient, θ j is the sliding average coefficient, p is the order of the stationary time series autoregressive model, and q is the order of the moving average data model.

Claims

1. A method for frequency conversion optimization control of circulating water of a seawater direct cooling unit considering tides, characterized in that: include: Step 100: Divide the circulating water pump combination mode according to the number of thermal power units, and divide the water level of the front water tank of the circulating water system into equal intervals according to the maximum and minimum values ​​of the seawater tide level at the power plant location within one year; Step 200: Establish a mathematical model of the circulating pump when the variable frequency pump runs alone and in parallel with the fixed speed pump, divide the water level correction pipeline characteristic curve of the front water tank of the circulating water system according to the previous step, divide the circulating pump speed ratio into equal intervals, calculate the front water tank water level and the pump group working point under different combination modes, that is, the total flow rate of the circulating water jellyfish pipe and the total power consumption of the circulating pump under a given circulating pump speed ratio, and use the least squares method to fit the functional relationship formula of the total flow rate of the circulating water jellyfish pipe and the total power consumption of the circulating pump with the circulating pump speed ratio; Step 300: Establish a cold end optimization mathematical model for coupling the condenser and the low-pressure cylinder of the steam turbine. Based on the theory of slightly increased output, the circulating water volume that maximizes the difference between the increased power of the steam turbine and the power consumption of the circulating pump is called the optimal circulating water volume. Determine the dividing line between the unit load and the circulating water inlet water temperature under different circulating water pump combination modes at each front water tank water level; Step 400: Taking the total power consumption of the pump group as the objective function of the circulating pump optimization energy consumption model, the target frequency of the variable frequency pump when different circulating water pump combination modes achieve the best vacuum at each front water tank water level is calculated based on the Bayesian optimization algorithm, and a multivariate regression model is constructed. The maximum likelihood method is used to fit the high-order polynomial function of the target frequency of the variable frequency pump and the unit load and the circulating water inlet water temperature; Step 500: Based on the high-order polynomial function obtained by the above offline optimization, a control strategy algorithm is written in the DCS. The algorithm input is the water level of the front water tank, the unit load and the circulating water inlet water temperature. The output is the circulating water pump combination mode and the target frequency of the variable frequency pump. The maximum variable frequency range of the variable frequency circulating water pump is set. The target frequency of the variable frequency pump calculated by the DCS is smoothed and filtered, and the amplitude is limited and the change rate is limited. Step 600: Based on the ARIMA model, the water level of the front water tank, the unit load and the circulating water inlet temperature in the future specified time period are predicted, and the DCS control strategy is combined with the prediction model to perform the pump cutting operation. The pump cutting operation is performed only when the input parameters calculated based on the cold end optimization mathematical model and the ARIMA model prediction show that there is no pump cutting operation in the future specified time period, and the DCS real-time control strategy determines that the water pump start-stop switching operation needs to be performed at the same time.

2. The method for frequency conversion optimization control of circulating water of a seawater direct cooling unit considering tides according to claim 1 is characterized in that: When the variable frequency pump is in operation, different speed ratios correspond to different flow rates and pump power consumption. When multiple pumps are operated in parallel, there are multiple combinations corresponding to multiple working points. The pumps at each working point have their own speed ratios to meet the total flow rate requirements. There is a working point that minimizes the total power consumption of the pump group, which is the objective function of the circulating pump optimization energy consumption model. The speed ratio corresponding to the working point at this time is the target speed ratio, and the corresponding frequency is the target frequency.

3. The method for frequency conversion optimization control of circulating water of a seawater direct cooling unit considering tides according to claim 1 is characterized in that: The circulating water pump combination mode is: when the unit-type unit circulating water system adopts a combination of 1 machine and 2 pumps, that is, 1 fixed speed pump and 1 variable frequency pump, when 2 thermal power units are running, the circulating water system has three situations: 2 pumps running, 3 pumps running and 4 pumps running, which can be specifically divided into: (1) Two fixed speed pumps run in parallel; (2) One fixed speed pump and one variable frequency pump run in parallel; (3) Two variable frequency pumps run in parallel; (4) Two variable-speed pumps and one fixed-speed pump run in parallel; (5) Two variable speed pumps and two fixed speed pumps operate in parallel.

4. The method for frequency conversion optimization control of circulating water of a seawater direct cooling unit considering tides according to claim 1 is characterized in that: During offline optimization, the circulating water temperature division interval is at least 1°C. When the measured temperature is not at the divided temperature point, the interpolation method can be used to obtain the variable frequency pump target frequency corresponding to the measured temperature from the variable frequency pump target frequency corresponding to the adjacent temperature points.

5. The method for frequency conversion optimization control of circulating water of a seawater direct cooling unit considering tides according to claim 1 is characterized in that: The optimization results obtained by offline simulation modeling calculation need to be fitted to obtain the polynomial function corresponding to the three-dimensional surface of water level, unit load and circulating water inlet temperature to simplify the DCS configuration workload. The circulating water optimization control strategy runs periodically in the DCS, and the DCS and the inverter are connected by hard wiring to issue control instructions.

6. The method for frequency conversion optimization control of circulating water of a seawater direct cooling unit considering tides according to claim 1 is characterized in that: The ARIMA model is used to make short-term predictions of key system parameters and provide confidence interval prediction results at different confidence levels combined with deterministic prediction results to form probabilistic prediction results, which are calculated using the following formula: where Y t is the predicted value, c is a constant, ε t and ε t-j is white noise, is the autoregressive coefficient, θ j is the sliding average coefficient, p is the order of the stationary time series autoregressive model, and q is the order of the moving average data model.

Images