TWI738094B - Method of controlling cooling water flow rate by using pattern recognition based on algorithms of inner product and gravity center - Google Patents

Abstract

A method is provided to control the flow rate of cooling water. The present invention uses pattern recognition, which is based on algorithms of inner product and gravity center. The method comprises the following steps: A historical operating database of a cooling water system is established. By using a thermodynamic model of cooling water, the historical operating database is verified whether being operated in a best condition. The verified information are used as samples with features. Whenever a new set of real-time operating data is obtained, an algorithm of tensor inner product is used to identify a number of the samples having similar features. Then, an algorithm of gravity center is used to find a number of gravity values of the samples having the similar features. What follows is to use the gravity values as control parameters to control flow rate of cooling water. Thus, the cooling water system is maintained to be operated at a historical best condition with energy and electricity saved.

Description

Cooling water flow control method using pattern recognition of tensor inner product and center of gravity

The present invention relates to a method of cooling using the pattern recognition of the inner product of the tensor and the center of gravity. Water flow control methods, especially those related to maintaining the cooling water system running in the best condition, to achieve the purpose of energy and power saving, especially the ability to optimize the operation of the condenser, to achieve the improvement of the power generation efficiency of coal-fired power plants, and indirectly achieve the reduction of carbon dioxide and carbon dioxide. The emission of PM2.5 fine suspended particulates, and the reduction of the emission of nitrogen oxides and sulfur dioxide to the environment.

As the economy grows, the demand for energy and electricity has also increased significantly; although natural Gas, petroleum, nuclear energy and renewable energy can be used for power generation. However, due to the wide range, availability, low price and high energy content of coal, it is still the main power supply in my country and the world. According to a 2018 report by the International Energy Agency, coal power generation accounts for about 38% of global electricity, but the carbon dioxide emissions produced account for more than two-thirds of the carbon dioxide produced by the entire power generation industry, leading to the greenhouse gas effect may become more and more serious The more serious it is, the irreversible climate change will be produced. According to Taipower’s statistics, coal-fired power generation accounted for about 40% in 2018. However, with the impact of PM2.5 fine suspended particulates and the rising public awareness of environmental protection, reducing carbon dioxide emissions has become a consensus among the people. The government has to build new coal-fired power plants. Not too possible.

As the steam in coal-fired power plants undergoes the Rankine cycle to convert heat energy into electrical energy, the steam The steam boiler is the heat source, and the condenser is used to dissipate heat. The performance of the two is critical to the power generation efficiency of the entire power plant. In view of this, some conventional technologies use a proportional-integral controller to control the variable-frequency cooling water pump by feedback of the cooling water temperature. The disadvantage is that it is only applicable to variable-frequency pumps, and the control parameters need to be set in When the cooling water pump is carried out in the state of shutdown, it cannot dynamically adjust itself. When the performance of the cooling water pump increases with the use of time, it may cause poor cooling water flow control. There are also some conventional technologies that use a neural network to build a model to predict the cooling water flow. However, the model needs to be trained, tested, and tuned, all based on experience.

Also check the published patents abroad, most of which use thermodynamics, heat transfer or aerodynamics To redesign the condensing system or cooling water tower, in addition to costly, it also needs to stop the operation to cooperate with the construction, which will also cause production losses; a small number of patents use finite element analysis to optimize operations, except for construction In addition to the time-consuming problem of modeling, it must also be realized by high-speed computing computer software and hardware. Therefore, general users cannot meet the needs of users in actual use.

The main purpose of the present invention is to overcome the above-mentioned problems encountered by the prior art and to improve Provides a cooling water flow control method for the condenser of a thermal power plant. It can use the establishment and correction of the historical operation database without adding or modifying any power plant facilities, and cooperate with the tensor inner product and center of gravity algorithm to find the best The cooling water flow rate can be used to optimize the operation of the cooling water system, which can effectively save the energy consumption of the cooling water pump. Energy saving and power saving can be achieved without additional equipment expenditures. The use of the tensor inner product and the pattern identification of the center of gravity for the cooling water flow Control Method.

The secondary objective of the present invention is to provide a system that can maintain the cooling water system running at the maximum To achieve the goal of energy saving and power saving, and to optimize the operation of the condenser, to improve the power generation efficiency of coal-fired power plants, to indirectly reduce the emission of carbon dioxide and PM2.5 fine suspended particulates, and to reduce the effects of nitrogen oxides and sulfur dioxide The cooling water flow control method is based on the model identification of the inner product of the tensor and the center of gravity for the discharge of the environment.

In order to achieve the above purpose, the present invention is a pattern recognition using tensor inner product and center of gravity Carrying out the cooling water flow control method, including the establishment of a historical operating data database of the cooling water system, and verifying whether the historical operating data is operating in the best state by the thermodynamic model of the cooling water system, and then using the verified data as a characteristic sample ; Whenever a new set of real-time operating data is obtained, it is convenient to use the tensor inner product algorithm to find several similar feature samples, and then use the center of gravity algorithm to find the center of gravity values of several similar feature samples, and use this center of gravity value as The control parameters of the cooling water pump control the flow of cooling water and maintain the operation of the cooling water system in the best historical state to achieve the purpose of energy saving and power saving.

The embodiments of the present invention are discussed in detail below. However, it is understandable that the embodiment provides Provides many applicable concepts, which can be implemented in a variety of specific content. The discussed and disclosed embodiments are for illustrative purposes only, and are not intended to limit the scope of the present invention.

Please refer to "Figure 1 to Figure 3", which are the use of tensor inner product in the present invention. Schematic diagram of the flow of the cooling water flow control method based on the pattern recognition of the center of gravity, and a preferred embodiment of the present invention A schematic diagram of the historical feature sample creation process and a schematic diagram of the pattern recognition process of a preferred embodiment of the present invention. As shown in the figure: the present invention is a method for cooling water flow control using tensor inner product and center of gravity pattern recognition, which includes the following steps: Historical feature sample establishment step s11: As shown in Figure 2, firstly, in step s111, collect the historical operating data of the cooling water system, and establish a historical operating data database in step s112, including the vapor pressure, temperature, flow rate, and steam pressure at a certain time. Enthalpy, entropy, etc., as well as cooling water inlet temperature, outlet temperature, flow rate, cooling water pump voltage, current, etc. Then in step s113, the historical operation data is imported into the thermodynamic model of the cooling water system for verification, and in step s114, it is judged whether the verification is operating in the best state; if so, go to step s115 to set the data as a characteristic sample and save it back to the historical operation Data database; if not, go to step s116 to change the cooling water flow rate, and then return to step s113 to import the thermodynamic model until step s114 determines that it meets the best operating state, and then step s115 to save it back to the historical operating data database As a historical feature sample. Pattern recognition step s12: As shown in Figure 3, first, in step s121, input a new set of real-time operating data, and in step s122 use the tensor inner product algorithm (Equation 1) to obtain the real-time operating data and all feature samples Then, in step s123, search for the first n feature samples with larger cosine values, and then use the center of gravity algorithm (formula 2) in step s124 to calculate the relative cosine value weight of the cooling water flow, and finally in step s125 according to the pump The Pu performance curve knows the control power value.

The real-time operating data in Figure 3 above can be stored in the history following the process in Figure 2 Run the data database and create a feature sample for the pattern identification in the second step.

（公式1） 其中，

為即時運轉數據，

為壓力、溫度與流量，

為某時刻歷史特徵樣本，

為壓力、溫度與流量。The calculation formula of the above tensor inner product algorithm is as follows:

(Formula 1) where,

For real-time operating data,

Is pressure, temperature and flow rate,

Is a sample of historical features at a certain moment,

Is pressure, temperature and flow.

（公式2） 其中，

為n個搜尋出特徵樣本之冷卻水流量，

為餘弦值。The calculation formula of the above-mentioned center of gravity algorithm is as follows:

(Formula 2) where,

Find out the cooling water flow rate of n characteristic samples,

Is the cosine value.

The present invention discloses a cooling water flow control method for the condenser of a thermal power plant, with the purpose of Without adding or modifying any power plant facilities, use the establishment and correction of the historical operation database, and cooperate with the tensor inner product and center of gravity algorithm to find the best cooling water flow rate to optimize the operation of the cooling water system. Effectively save the energy consumption of the cooling water pump, and achieve the purpose of energy saving and electricity saving without adding additional equipment expenditure. Its technical characteristics are as follows: (1) The calculation method is simple and does not require high-end computing equipment; (2) No need to add additional equipment expenditure; and (3) There is no problem of control parameter setting and adjustment.

As a result, one of the effects that the present invention can produce is to maintain the cooling water system running at In the best state, achieve the purpose of energy saving and power saving. The technical verification of the present invention is currently being carried out with actual power plant operating data, and it is estimated that the electricity can be saved by 15.2%. The second effect that the present invention can produce is that it can optimize the operation of the condenser, improve the power generation efficiency of coal-fired power plants, indirectly reduce the emission of carbon dioxide and PM2.5 fine suspended particulates, and reduce the effects of nitrogen oxides and sulfur dioxide. Environmental emissions.

In summary, the present invention uses the pattern recognition of the inner product of the tensor and the center of gravity to perform cooling The cooling water flow control method can effectively improve the various shortcomings of the conventional use, maintain the cooling water system in the best condition, achieve the purpose of energy saving and power saving, and can optimize the operation of the condenser to improve the power generation efficiency of coal-fired power plants, indirectly. Reduce the emission of carbon dioxide and PM2.5 fine suspended particles, and reduce the emission of nitrogen oxides and sulfur dioxide to the environment, so that the production of the present invention can be more advanced, more practical, and more in line with the needs of users. It is indeed in line with the invention. For the requirements of a patent application, a patent application shall be filed in accordance with the law.

However, the above are only the preferred embodiments of the present invention and should not be limited by this The scope of implementation of the present invention; therefore, all simple equivalent changes and modifications made in accordance with the scope of the patent application of the present invention and the content of the description of the invention should still fall within the scope of the patent of the present invention.

s11～s12: steps s111～s115: steps s121～s125: steps

Figure 1 shows the flow of the cooling water flow control method using the pattern recognition of the tensor inner product and the center of gravity according to the present invention. 程Schematic diagram. Figure 2 is a schematic diagram of the historical feature sample creation process of a preferred embodiment of the present invention. Figure 3 is a schematic diagram of the pattern recognition process of a preferred embodiment of the present invention.

s11~s12: steps

Claims (5)

A method of cooling water flow control using the pattern recognition of the inner product of the tensor and the center of gravity, The system includes the following steps: Step 1: Establish a database of historical operating data of the cooling water system, and then import the historical operating data into the thermodynamic model of the cooling water system to verify whether it is operating in the best state. If it is, use the data as a characteristic sample. If not, , After changing the cooling water flow rate, import the thermodynamic model until it meets the best operating state, and save it back to the historical operating data database as a historical feature sample; and Step 2: Input a new set of real-time operation data, use the tensor inner product algorithm to find the cosine value of the real-time operation data and all feature samples, search for several feature samples with larger cosine values, and then use the center of gravity algorithm Calculate the relative cosine value weighted cooling water flow, and finally get the control power value according to the pump performance curve. According to the first item of the scope of patent application, the pattern recognition using tensor inner product and center of gravity A cooling water flow control method, wherein the step one is to establish a historical operating data database of the cooling water system, including the vapor pressure, temperature, flow, enthalpy and entropy at a certain moment, as well as the cooling water inlet temperature, outlet temperature, flow rate, etc. Cooling water pump voltage and current. The cooling water flow control method using the pattern recognition of the tensor inner product and the center of gravity described in item 1 of the scope of patent application, wherein the calculation formula of the step two tensor inner product algorithm is as follows:

, in,

For real-time operating data,

Is pressure, temperature and flow rate,

Is a sample of historical features at a certain moment,

Is pressure, temperature and flow. The cooling water flow control method using the pattern identification of the inner product of the tensor and the center of gravity described in item 1 of the scope of patent application, wherein the calculation formula of the two-center-of-gravity algorithm in this step is as follows:

, in,

Find out the cooling water flow rate of n characteristic samples,

Is the cosine value. According to the first item of the scope of patent application, the pattern recognition using the inner product and the center of gravity of the tensor is used. A cooling water flow control method is implemented, wherein the real-time operation data of the step 2 can be stored in the historical operation data database according to the process of the step 1, and a characteristic sample is created for the mode identification of the subsequent step 2.

Images