CN115074776B - Intelligent Adaptive Control System and Method for Hydrogen Production by Electrolysis of Water Adapting to Wide Power Fluctuation - Google Patents

Abstract

The invention relates to an intelligent self-adaptive control system and method for preparing hydrogen by electrolyzing water, which is suitable for wide power fluctuation, wherein the system comprises: expert experience knowledge base module: the method comprises the steps of obtaining expected output parameters of the electrolytic cell under the condition that the limit of boundary conditions is met; feedback compensation module: the method comprises the steps of detecting a deviation value between an output parameter and a boundary condition after the operation of the electrolytic cell, and outputting an expert experience compensation value; and (3) a water electrolysis hydrogen production module: the method is used for controlling the stable operation of the electrolytic tank under the condition that the input power fluctuates according to the expert experience compensation value and the output value of the expert experience knowledge base. According to the invention, under the premise of adapting to wide power fluctuation, the parameter set value of the hydrogen production system can be obtained through the cooperation of each module, so that the electrolytic water hydrogen production module can safely and stably operate under the influence of renewable energy fluctuation, the hydrogen production efficiency is improved, and the prepared hydrogen is high in purity.

Description

Intelligent adaptive control system and method for hydrogen production by water electrolysis adapting to wide power fluctuations

Technical Field

The invention relates to the technical field of water electrolysis hydrogen production, and in particular to an intelligent adaptive control system and method for water electrolysis hydrogen production adapting to wide power fluctuations.

Background Art

As a clean and low-carbon energy, hydrogen energy has various advantages such as cleanliness, storage and high energy carrier, and is considered to be the most promising secondary energy in the 21st century. It is also one of the key paths to promote energy structure change and achieve the "dual carbon" goal. However, at present, hydrogen production from fossil energy is the mainstream method of hydrogen production, and the purity of the produced hydrogen is relatively low. In order to reduce the harm of large-scale use of fossil fuels to human health and the environment, the country encourages the use of renewable energy to produce hydrogen. The hydrogen produced in this way is green hydrogen, and the purity is as high as 99.95% or more. Green hydrogen will be the mainstream in the future.

Among them, in the process of hydrogen production, alkaline water electrolysis is the most promising method for industrial production of green hydrogen. However, due to the uncertainty of renewable energy such as wind and light power generation, the given power of the electrolyzer will have large fluctuations, which may cause some key parameters in the electrolyzer system to exceed the safety boundary conditions, resulting in serious consequences. For example, at a certain moment, the input power of the electrolyzer system fluctuates over a large range, which will cause the hydrogen content in the oxygen in the oxygen scrubber to be out of the safe range (below 2%). In order to avoid the risk of explosion, the system will automatically alarm and shut down the electrolyzer system, further affecting the hydrogen production efficiency and the safety of the system. Therefore, it is of great significance to study the control method of the electrolyzer system that adapts to wide power fluctuations, and to ensure the continuous and stable operation of the electrolyzer and maximize the hydrogen production under the premise of safe operation of the system.

Summary of the invention

The purpose of the present invention is to provide an intelligent adaptive control system and method for hydrogen production by water electrolysis that can adapt to wide power fluctuations. Through the expert experience knowledge base model, the mutual cooperation between the feedback compensation module and the water electrolysis hydrogen production module, the electrolyzer can adapt to different power fluctuations, ensure the safe and stable operation of the electrolyzer, and improve the hydrogen production efficiency of the electrolyzer.

To achieve the above object, the present invention provides the following solutions:

Intelligent adaptive control system for hydrogen production by water electrolysis that adapts to wide power fluctuations, including:

Expert experience knowledge base module: used to obtain the expected output parameters of the electrolytic cell under the constraints of boundary conditions, wherein the boundary conditions are used to limit system parameters to ensure the safe operation of the system, and the expected output parameters are used to make the parameters of the system during operation meet the set boundary conditions;

Feedback compensation module: used to detect the deviation between the output parameter of the electrolytic cell after operation and the boundary condition, and output the expert experience compensation value;

Water electrolysis hydrogen production module: used to control the stable operation of the electrolyzer when the input power fluctuates according to the expert experience compensation value and the output value of the expert experience knowledge base.

Preferably, the expert experience knowledge base module includes:

Expert experience knowledge base model: used to obtain the expected output parameters of the electrolytic cell under the constraint of the boundary conditions according to the current input power of the electrolytic cell and the current measured operation parameter Y _k of the electrolytic cell.

Preferably, the measured operating parameters _Yk of the electrolytic cell include: system pressure, alkali solution flow rate, electrolytic cell temperature, liquid level difference between the oxygen separator and the hydrogen separator, hydrogen content in oxygen in the oxygen scrubber, and oxygen content in hydrogen in the hydrogen scrubber.

Preferably, the expert experience knowledge base module also includes a query and matching unit, which is used to query and match the expert experience knowledge in the knowledge base that meets the boundary conditions by calculating the current operating characteristic parameters of the electrolytic cell based on a multi-attribute similarity algorithm, and screen out the expected output value in the knowledge base that is closest to the current operating condition of the electrolytic cell.

Preferably, the multi-attribute similarity algorithm includes a nearest neighbor algorithm, Euclidean distance and structural similarity, and the overall similarity between the current operating parameters of the electrolytic cell and the expert experience knowledge is obtained based on the multi-attribute similarity algorithm, and query and matching are performed based on the similarity.

Preferably, the expert experience knowledge base module also includes an evaluation and correction unit and a storage and addition unit. The evaluation and correction unit is used to evaluate and correct the reuse results of the expert experience knowledge in the expert experience knowledge base, and judge whether to correct the expert experience knowledge based on whether the important parameters during the operation of the electrolytic cell meet the boundary conditions. If the current important parameters exceed the boundary conditions, the expert experience knowledge needs to be corrected. If the important parameters are all within the boundary conditions, the expert experience knowledge does not need to be corrected; wherein the important parameters include hydrogen content in oxygen, oxygen content in hydrogen, liquid level difference and alkali solution temperature; the storage and addition unit is used to add new expert experience knowledge.

Preferably, the feedback compensation module includes:

Expert rule establishment unit: used for setting rules according to the deviation between the important parameters output after the operation of the electrolytic cell and the boundary value, and the expert gives the correlation coefficient of parameter compensation for each deviation according to experience and rules; wherein the rules include: the error between the hydrogen content in oxygen during the operation of the electrolytic cell and the boundary value threshold, the error between the oxygen content in hydrogen during the operation of the electrolytic cell and the boundary value threshold, the error between the liquid level difference during the operation of the electrolytic cell and the boundary value threshold, and the error between the alkali solution temperature during the operation of the electrolytic cell and the boundary value threshold;

Inference engine unit: used to infer the corresponding expert experience compensation value according to the difference between the important parameters output during the operation of the electrolytic cell and the boundary value, using an exhaustive item-by-item search algorithm through expert experience rules, and output the expert experience compensation value.

Preferably, the water electrolysis hydrogen production module is used to perform transient PID control on the pressure regulating valve and the pneumatic regulating valve in the electrolyzer through a combination of fuzzy rules and neural networks, to ensure that the current pressure of the electrolyzer, the liquid levels of the oxygen separator and the hydrogen separator, and the alkali solution temperature and flow rate respond to the given values of important parameters.

An intelligent adaptive control method for hydrogen production by water electrolysis adapting to wide power fluctuations includes:

Constructing an expert experience knowledge base model, and obtaining the expected output parameters of the electrolyzer system under the constraints of boundary conditions according to the current input power and measured parameters of the electrolyzer, wherein the boundary conditions include: hydrogen content in oxygen, oxygen content in hydrogen, alkali solution flow rate, alkali solution temperature, and liquid level difference between the oxygen separator and the hydrogen separator, and the expected output parameters include a given value of system pressure, a given value of alkali solution temperature, and a given value of alkali solution flow rate;

Running the electrolytic cell system based on the expected output parameter, detecting the deviation between the output parameter of the electrolytic cell system after running and the boundary condition through a feedback compensation model, and outputting an expert experience compensation value;

According to the expert experience compensation value and the expected output parameters, the pressure regulating valve and the pneumatic regulating valve in the electrolytic cell are controlled by combining fuzzy rules and neural networks to perform transient PID control. At the same time, the output boundary important parameter detection values are sent to the feedback compensation model and the expert experience knowledge base model for monitoring and correction at the same time.

The beneficial effects of the present invention are:

The present invention can adapt to the wide power fluctuation of the alkaline electrolyzer caused by the uncertainty of renewable energy generation, and adaptively control the given parameters of the electrolyzer (system pressure, alkali solution flow rate, alkali solution temperature) through an intelligent control method to keep the important parameters of the electrolyzer system (such as hydrogen content in oxygen) within a safe range, thereby effectively improving the hydrogen production efficiency and purity of the electrolyzer while ensuring the safe and stable operation of the electrolyzer system.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

FIG1 is a schematic diagram of the module structure of an intelligent adaptive control system for hydrogen production by electrolysis of water that is adaptable to wide power fluctuations according to an embodiment of the present invention;

FIG2 is a line graph of hydrogen content in oxygen under different electrolyzer input power ranges and different system pressures according to an embodiment of the present invention;

FIG3 is a schematic diagram of a safe area of hydrogen content in oxygen under different power ranges and different system pressures according to an embodiment of the present invention;

FIG4 is a line graph of oxygen content in hydrogen under different power ranges and different system pressures according to an embodiment of the present invention;

FIG5 is a schematic diagram of the liquid level difference of the separator under different power ranges and different system pressures according to an embodiment of the present invention;

FIG6 is a schematic diagram of energy consumption and energy efficiency under different power pressures according to an embodiment of the present invention;

7 is a flow chart of an intelligent adaptive control method for hydrogen production by electrolysis of water adapted to wide power fluctuations according to an embodiment of the present invention;

FIG8 is a schematic diagram of the workflow of the expert experience knowledge base module according to an embodiment of the present invention.

DETAILED DESCRIPTION

The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

After renewable energy generates electricity, it outputs power to the alkaline electrolyzer through the DC microgrid. The power supplied to the electrolyzer fluctuates greatly, which may cause important parameters of the hydrogen production system to exceed the safe operation boundary conditions. For example:

1) The hydrogen content in the oxygen in the oxygen scrubber exceeds 2%. Once it exceeds 2%, there is a risk of explosion. In addition, the hydrogen produced by alkaline water electrolysis requires high purity, which limits the oxygen content in the hydrogen in the hydrogen scrubber to no more than 0.5%;

2) If the liquid level of the oxygen separator and the liquid level of the hydrogen separator deviate greatly, alkali liquid may enter the scrubber and be sprayed out from the vent through the scrubber. If someone passes under the air defense port, an alkali burn accident may occur. In addition, if the liquid level on one side is too low, the gas and alkali liquid in the separator may enter the circulation pump at the same time, causing a large fluctuation in the circulation volume of the alkali liquid, or even stop rotating. If the alkali liquid stops circulating, the liquid level deviation may continue to increase, and the gas in the separator on one side will enter the separator on the other side, and hydrogen and oxygen will mix during separation, which is very likely to explode in the separator, causing serious safety accidents. Therefore, the liquid level difference between the oxygen separator and the hydrogen separator should not exceed 5cm;

3) During the operation of the alkaline electrolytic cell, the cell temperature is mainly controlled by alkali solution, in which the alkali solution temperature is required to be controlled at around 65°C, with fluctuations of less than 1.1°C. The greater the alkali solution flow rate, the lower the hydrogen cell temperature and oxygen cell temperature. Since the diaphragm between the cathode and anode of the electrolytic cell has certain temperature requirements, the hydrogen cell temperature and oxygen cell temperature should be less than 85°C during the operation of the electrolytic cell. If the alkali solution flow rate is too low, causing the temperature of the separator to exceed the boundary temperature, the diaphragm between the cathode and anode of the electrolytic cell will be damaged, causing the hydrogen and oxygen on both sides to mix, resulting in serious consequences.

4) In addition, the alkali liquid flow rate will also affect the purity of hydrogen and oxygen. If the alkali liquid flow rate is too large, the gas-liquid mixture produced by electrolysis in the separator will carry more impurities, resulting in a decrease in the purity of oxygen and hydrogen in the oxygen separator and hydrogen separator. In this embodiment, the alkali liquid flow rate is set within a range of ^3.0-4.5m3 /h.

In addition, if the above-mentioned hydrogen content in oxygen, oxygen content in hydrogen, liquid level difference between oxygen separator and hydrogen separator, alkali liquid flow rate, and alkali liquid temperature exceed the boundary conditions, it will not only have a serious impact on system safety but also lead to a decrease in hydrogen production efficiency and hydrogen output.

Referring to FIG. 1 , this embodiment provides an intelligent adaptive control system for hydrogen production by electrolysis of water that is adaptable to wide power fluctuations, including:

Expert experience knowledge base module: used to obtain the expected output parameters of the electrolytic cell under the constraints of boundary conditions, wherein the boundary conditions include: hydrogen content in oxygen, oxygen content in hydrogen, alkali solution flow rate, alkali solution temperature and liquid level difference between oxygen separator and hydrogen separator, and the expected output parameters include a given value of system pressure, a given value of alkali solution concentration and a given value of alkali solution flow rate;

Feedback compensation module: used to detect the deviation between the output parameter of the electrolytic cell after operation and the boundary condition, and output the expert experience compensation value;

Water electrolysis hydrogen production module: used to control the stable operation of the electrolyzer when the input power fluctuates according to the expert experience compensation value and the output value of the expert experience knowledge base.

(1) Expert Experience Knowledge Base Model

The expert experience knowledge base model is based on the power input to the current electrolyzer system, comprehensively considering the current operating measured parameters Y _k of the current electrolyzer system, and under the restriction of the boundary condition B _k , according to the expert experience knowledge base model, coordinately giving the closed-loop given value Z _k of the electrolyzer system under the restriction of satisfying the boundary condition.

The measured parameters (Y _k ) of the electrolyzer system input into the expert experience knowledge base are system pressure, alkali solution flow rate, electrolyzer temperature, liquid level difference between oxygen separator and hydrogen separator, hydrogen content in oxygen in oxygen scrubber, and oxygen content in hydrogen in hydrogen scrubber. Under the constraint of boundary conditions, the expected output parameters of the electrolyzer system are obtained by combining knowledge base reasoning and mathematical model, which are system pressure given value, alkali solution temperature given value, and alkali solution flow rate given value.

The steps are as follows:

① Establish a basic knowledge base for water electrolysis hydrogen production system

Through a large number of experiments, when the electrolyzer inputs different power ranges (20%, 40%, 60%, 80%, 100%), by changing the system pressure, alkali solution temperature, alkali solution flow parameters, the electrolyzer system can operate stably under the restriction of boundary conditions, where the boundary conditions are that the hydrogen content in oxygen in the oxygen scrubber is less than 2% (50% LFL), the oxygen content in hydrogen in the hydrogen scrubber is less than 0.5%, and the liquid level difference between the oxygen separator and the hydrogen separator is within 5cm, the alkali solution flow is within ^3-4.5m3 /h, the alkali solution temperature is maintained at about 65℃, the upper and lower fluctuation range is maintained at 1.1℃, the electrolyzer temperature is less than 85℃, etc. Figures 2 to 6 are obtained through experiments, and their representative meanings are introduced in turn below.

Figure 2 is a graph showing the change in hydrogen content in oxygen when the electrolyzer is input with different power ranges and by changing different system pressures. It can be seen that when the electrolyzer is input with different power ranges, at the current system pressure (the pressure in the oxygen separator), if the hydrogen content in oxygen exceeds the safe range by 2% (50% LFL), the hydrogen content in oxygen can be brought back to the safe range by appropriately changing the system pressure.

Under each power test, the electrolyzer runs continuously for more than two hours to ensure that the electrolyzer can operate for a long time under this power. For example, when the input power of the electrolyzer is 40% of the rated power, if the system pressure is 1.6Mpa, the hydrogen content in oxygen in the oxygen scrubber has exceeded 2%. By changing the system pressure to 1.0Mpa, it can be seen in Figure 2 that the hydrogen concentration in oxygen drops below 2%, and the system returns to the safe range. Therefore, when the hydrogen concentration in oxygen is not within the safe range, the safe and stable operation of the system can be guaranteed by appropriately changing the pressure in the system, and based on this, the safe area of hydrogen concentration in oxygen under different input powers and different system pressures of the electrolyzer is established.

As shown in Figure 3, when the LFL is 50%, the red area is the area where the hydrogen content in oxygen is within the safe range; when the LFL is 75%, the red and blue areas are the areas where the hydrogen content in oxygen is within the safe range; Area 3 is the undetectable area; Area 4 is the unsafe area; and Area 5 is the temporarily available area.

Figure 4 shows the change of oxygen content in hydrogen in the hydrogen scrubber when the electrolyzer inputs different power ranges and changes the system pressure. It can be found from Figure 4 that the concentration of oxygen in hydrogen can be well maintained within a safe range. When the electrolyzer input power is determined, the oxygen content in hydrogen in the hydrogen scrubber can be reduced by changing the system pressure. In order to increase the concentration of hydrogen produced by the electrolytic water hydrogen production system, when the oxygen content in hydrogen in the electrolyzer system exceeds the hydrogen purity requirement (<0.5%), the purpose of adjusting the oxygen content in hydrogen can be achieved by adjusting the system pressure, thereby improving the purity of hydrogen produced by the system.

Figure 5 shows the change between the liquid level difference of the oxygen separator and the hydrogen separator under different power ranges and pressures of the electrolyzer input. The liquid level difference of the hydrogen-oxygen separator of the water electrolysis hydrogen production system is required to be close to zero, and the deviation does not exceed 0.7 cm. The maximum liquid level deviation shown in Figure 5 does not exceed 0.7 cm, which meets the requirements. Therefore, it is concluded that the input power of the electrolyzer and the system pressure have no obvious correlation with the liquid level difference of the hydrogen-oxygen separator, so the safe range of hydrogen content in oxygen can be met by adjusting the system pressure at different powers. Figure 6 shows the power and energy efficiency consumed by the system under different powers and system pressures. With the increase of power, the energy consumption of the stack increases while the energy efficiency decreases; and the change of system pressure has no obvious effect on these two indicators, so it also shows that the safe range of hydrogen content in oxygen can be met by adjusting the system pressure.

Tables 1 and 2 show the changes in oxygen tank temperature and hydrogen tank temperature in the electrolyzer under different alkali solution flow rates, as well as the changes in hydrogen purity and oxygen purity in the separator. It can be seen from Table 1 that the greater the alkali solution flow rate, the lower the oxygen tank temperature and hydrogen tank temperature. When the alkali solution flow rate is 2.6 ^{m 3} /h, the tank temperature has exceeded 85°C, which is not within the safe range of the electrolyzer operating temperature, and is not conducive to the safe operation of the electrolyzer system. In Table 2, as the alkali solution flow rate increases, the oxygen purity and hydrogen purity in the oxygen separator and hydrogen separator decrease slightly, so the alkali solution flow rate can be controlled to achieve the effect of adjusting the electrolyzer temperature and the purity of oxygen and hydrogen. Therefore, the content of hydrogen in oxygen can also be changed by adjusting the alkali solution flow rate at different working powers, and it can also be controlled at 2% (50% LFL). In this way, the content of hydrogen in oxygen can be controlled not only by changing the system pressure, but also by combining the system pressure and the alkali solution flow rate to broaden the power adjustment range.

Table 1

Table 2

Based on this, the original expert experience knowledge base is established according to the experience gained from a large number of experiments. The expert experience knowledge base can be expressed as follows:

E _k ={T _k ,P _k ,Y _k ,B _k ,Z _k ,S _k }

In the formula, _Ek is the Kth expert experience knowledge (K = 1, 2, 3, ..., m, where m is the number of expert experience knowledge); _Tk is the storage time of the expert experience knowledge _Ek , _Pk is the ratio of the input power of the expert experience knowledge _Ek to the rated power, _Yk = { _yk1 , _yk2 , _yk3 , _yk4 , _yk5 , _yk6 } are the characteristic parameters of the electrolyzer system of the expert experience knowledge _Ek during operation, and _yk1 , _yk2 , ..., _yk6 respectively represent the system pressure, alkali solution flow rate, electrolyzer temperature, liquid level difference between oxygen separator and hydrogen separator, hydrogen content in oxygen, and oxygen content in hydrogen under the safe and stable operation of the electrolyzer system in the Kth expert experience knowledge. B _k = {b _k1 , b _k2 , b _k3 , b _k4 , b _k5 } is the boundary condition of the important parameters of the stable operation of the electrolyzer system in the expert experience knowledge E _k , b _k1 , b _k2 , b _k5 respectively represent the hydrogen content in oxygen in the oxygen scrubber < 2%, the oxygen content in hydrogen in the hydrogen scrubber < 0.5%, 3m ³ /h ≤ alkali solution flow ≤ 4.5m ³ /h, the liquid level difference between the oxygen separator and the hydrogen separator < 5cm, 63.9℃ ≤ alkali solution temperature ≤ 66.1℃. _{Z k} = {z _k1 , z _k2 , z _k3 } is the expected value of the output of the expert experience knowledge, z _k1 , z _k2 , z _k3 respectively represent the given value of the system pressure, the given value of the alkali solution temperature, and the given value of the alkali solution flow. S _k is the comprehensive similarity between the current characteristic parameters of the electrolyzer system and the Kth expert experience knowledge. The input and output of an expert experience knowledge E _k are shown in Table 3.

Table 3

Query and matching unit:

This embodiment uses multi-attribute similarity calculation to query and match expert experience knowledge in the knowledge base that meets the safety boundary conditions through the current electrolytic cell system operation characteristic parameters, and screen out the expected output value in the knowledge base that is closest to the current electrolytic cell operating conditions (as shown in Figure 8). First, the current electrolytic cell system situation is defined as expert experience knowledge E _n , P _n is the electrolytic cell power input ratio in the current situation, Y _n = {y _n1 , _yn2 , ..., _yn6 } is the various parameter values of the electrolytic cell system in the current situation, first, for the electrolytic cell input power attribute, the input power and all expert experience knowledge are calculated according to the optimized nearest neighbor algorithm, so as to determine the expert experience knowledge within the adjacent power range of the current electrolytic cell input power. The optimized nearest neighbor algorithm is as follows:

The exponential form can make the similarity calculation more accurate. sim(P _i,k ,P _i,n ) represents the similarity between the input power of the expert experience knowledge E _k and the input power of the current situation. Find the expert experience knowledge of the adjacent range of input power in the current situation. Among them, the various characteristic parameters of the electrolytic cell system may be missing or the characteristic attribute value may be 0, so the characteristic information is incomplete, which affects the similarity calculation. Therefore, structural similarity is added to reduce this impact.

Structural similarity only calculates the similarity of attributes whose values are not 0 between the current situation and the historical expert experience knowledge, which can effectively avoid the problem of incomplete information. Assuming P = {all non-empty attribute sets of the electrolytic cell system in the current situation}, Q = {all non-empty attribute sets in the historical expert experience knowledge Q}, the structural similarity S of P and Q is expressed as follows:

Among them, ω _∩ is the sum of the weight values of all attributes in the intersection of sets P and Q, and ω _∪ is the sum of the weight values of all attributes in the union of sets P and Q.

The Euclidean distance formula between the current parameters of the electrolyzer system and the previous expert experience knowledge is:

Where ω _i represents the characteristic weight value of expert experience knowledge. Through expert experimental experience, it can be concluded that the system pressure characteristics have a great influence on the important parameters of the electrolytic cell system. Under the condition of a given electrolytic cell power, under the current system pressure, the hydrogen content in oxygen in the electrolytic cell system is not within the safe range. By properly adjusting the system pressure, the hydrogen content in oxygen can be within the safe range, as shown in Figure 2. Compared with the system pressure, the alkali solution flow rate and system liquid level have little effect on the important parameters. Based on this, according to the degree of influence on the safe operation of the electrolysis system, ω _i is set to: ω _i = {0.9, 0.05, 0.05, 0, 0, 0}.

The nearest neighbor algorithm, Euclidean distance and structural similarity are combined to obtain the overall similarity between the current situation of the electrolyzer and the historical expert experience knowledge:

Assume that sim _max is the maximum value of the similarity between the current electrolyzer system situation and the historical expert experience knowledge, that is:

The comprehensive similarity threshold sim _yz can be set as:

The threshold Y _YZ is given by expert experience and is set to 0.9 here. All expert experience knowledge with overall similarity sim(E _n ,E _k ≥sim _yz ) is retrieved from the expert experience knowledge base, and then the solution {Z _k }, time T _k and overall similarity sim(E _n ,E _k ) of the expert experience knowledge are recorded, and then they are arranged in descending order according to the attribute values of "overall similarity" and "expert experience knowledge storage time", waiting for the next step of processing.

Reuse of expert experience and knowledge:

The expert experience knowledge with the maximum similarity sim _max is selected from the matched expert experience knowledge and its number Num is determined.

If Num＝1, it means there is only one expert experience knowledge with the maximum similarity. Let this expert experience knowledge be E _k , 1≤k≤m, let the next expert experience knowledge of the expert experience knowledge E _k in the matching expert experience knowledge data table be E _h , 1≤h≤m, because the matching expert experience knowledge is sorted in descending order according to the attribute values of "overall similarity" and "expert experience knowledge storage time" when it is retrieved, so E _h should be the second largest similarity and the one with the latest expert experience knowledge storage time. Let the expected output of expert experience knowledge E _h be Z _h , the overall similarity be sim _h , and the expected output of expert experience knowledge E _k be Z _k , then the expected output Z _hk of the description in the current situation is:

If Num>1, it means that there are multiple expert experience knowledge with the same maximum overall similarity. Suppose there are f expert experience knowledge. Assume that these expert experience knowledge _Ei , i=1…f are arranged in descending order according to the expert experience knowledge storage time attribute value: _E1 , _E2 … _Ef , _Z1 , _Z2 … _Zf are their corresponding expected outputs, then the expected output of the description in the current situation is:

Where θ _i is the time weighted coefficient of the expert experience knowledge reuse this time, satisfying θ ₁ ≥θ ₂ ≥…≥θ _l , which can be determined according to specific circumstances or experience.

Evaluation and Correction Unit:

In order to verify the effectiveness of the reuse results of expert experience knowledge, expert experience knowledge evaluation and correction must be performed. The first step is to evaluate the reuse results. If successful, no correction is required. Otherwise, expert experience knowledge correction is performed to improve the accuracy of the setting module. In this embodiment, the expert experience knowledge evaluation is based on the feedback of its running effect in the actual environment, and the expert experience knowledge correction is performed based on the problems in the execution process.

After the electrolyzer system is running, the hydrogen content in oxygen, the oxygen content in hydrogen, the liquid level difference and the alkali solution temperature will have a certain lag. Therefore, these four important parameters are tested every hour, and the four important parameter values are fed back to the expert experience knowledge base to determine whether the important parameters under the current operating state meet the boundary conditions. Whether to correct the expert experience knowledge is determined based on whether the boundary conditions are met. If the current important parameters exceed the boundary conditions, the expert experience knowledge needs to be corrected. For example, if the hydrogen content in oxygen in the oxygen scrubber after the electrolyzer system is running exceeds the limit of 2%, it will affect the system operation and the expert experience knowledge will be corrected. If the important parameters are all within the boundary conditions, there is no need to correct the expert experience knowledge.

Storage and Add-on Units:

When new expert experience knowledge is added to the historical expert experience knowledge base, the overall similarity between the new expert experience knowledge and all expert experience knowledge stored in the historical expert experience knowledge base is calculated. If all the similarities are less than or equal to a given threshold (the threshold is 0.8), the new expert experience knowledge is added. If there is at least one similarity greater than the threshold, it is not stored.

(2) Feedback Compensation Model

After the expected output parameters selected from the expert experience knowledge base are given to the electrolyzer, some parameter changes during the operation of the electrolyzer may have a serious impact on the safety of the system. The four most important parameters that have the greatest impact on system safety are the hydrogen content in oxygen in the oxygen scrubber, the oxygen content in hydrogen in the hydrogen scrubber, the liquid level difference between the oxygen separator and the hydrogen separator, and the alkali solution temperature. The hydrogen content in oxygen is strictly required to be below 2%, the oxygen content in hydrogen must be below 0.5%, the liquid level difference between the oxygen separator and the hydrogen separator cannot exceed 5cm, and the alkali solution temperature must be maintained at around 65°C.

In this embodiment, in order to deal with the situation where important parameters exceed the boundary conditions, a feedback compensation model is added to the intelligent control model. After the electrolyzer is operated, the four important parameters are tested every hour. The feedback compensation model is based on the deviation between the four important parameters detected after the electrolyzer is operated and the boundary conditions. The given values of each parameter of the intelligent control system are corrected through the set expert rules, so that the parameters affecting the system safety and hydrogen production purity return to the safe range.

The feedback compensation model is mainly divided into three parts:

Expert rule building unit:

Experts set rules for the deviations (△e1, △e2, △e3, △e4) between the four important parameters output after the operation of the electrolyzer and the boundary values through practical experience and a large number of experiments. The rule table is shown in Table 4. Among them, △e1, △e2, △e3, △e4 are the errors between the hydrogen content in oxygen and the boundary value of 2% during the operation of the electrolyzer, the error between the oxygen content in hydrogen and the boundary value of 0.5% during the operation of the electrolyzer, the error between the liquid level difference and the boundary value of 5cm during the operation of the electrolyzer, and the error between the alkali solution temperature and the boundary value of 65°C during the operation of the electrolyzer. △p, △f, △T respectively represent the system pressure compensation value, alkali solution flow compensation value, and alkali solution temperature compensation value obtained based on expert experience. k _1,1 , k _1,2 , ..., k _4,1 , k _4,2 respectively represent the correlation coefficients of the compensation of three given parameters of the electrolyzer given for each deviation based on the experience and rules of the experts for four different deviations.

Table 4

Inference engine unit:

The inference engine uses an exhaustive item-by-item search algorithm to infer the corresponding expert experience compensation value based on the difference between the four important parameters output during the operation of the electrolytic cell and the boundary value, and outputs the expert experience compensation value inferred by the inference engine.

Water electrolysis hydrogen production module:

The electrolysis water hydrogen production module uses the expert experience knowledge base model system and the feedback compensation model to correct the given parameters of the electrolyzer system, including the system pressure given value, the alkali solution temperature given value, and the alkali solution flow given value. Through the combination of fuzzy rules and neural networks, the pressure regulating valve and the pneumatic regulating valve in the electrolyzer system are transiently PID controlled to ensure that the pressure of the electrolyzer system, the liquid level of the oxygen separator and the hydrogen separator, and the alkali solution temperature and flow can quickly respond to the given values of the parameters, ensuring the safe and stable operation of the electrolyzer hydrogen production system under input power fluctuations. At the same time, after the electrolyzer is running, because the parameter changes will have hysteresis, the four parameters are detected every hour, and the detected four parameter values are output to the expert experience knowledge base model. These four parameters are the hydrogen content in oxygen in the oxygen scrubber, the oxygen content in hydrogen in the hydrogen scrubber, the liquid level difference between the oxygen separator and the hydrogen separator, and the alkali solution temperature. The reason for outputting these four parameters to the expert experience knowledge base model is to determine whether they are within the boundary range. If they are within the range, the system is running normally and the expert experience knowledge does not need to be corrected. If it is not within the boundary, the expert experience knowledge needs to be revised.

As shown in FIG. 7 , this embodiment also provides an intelligent adaptive control method for hydrogen production by electrolysis of water that is adaptable to wide power fluctuations, including:

The electrolysis water hydrogen production system uses the given parameters of the electrolyzer system corrected by the expert experience knowledge base model system and the feedback compensation model, and combines fuzzy rules and neural networks to perform transient PID control on the pressure regulating valve and pneumatic regulating valve in the electrolyzer system. At the same time, it outputs the detection values of four important boundary parameters to the feedback compensation model and the expert experience knowledge base model every hour. The purpose of outputting to the feedback compensation model is to derive the compensation values of the four important parameters according to the set expert experience rules based on the deviation between the four important parameters and the boundary conditions; the purpose of outputting to the expert experience knowledge base model is to determine whether the current parameter detection value is within the boundary range. If it is within the range, the system operates normally and the expert experience knowledge does not need to be corrected. If it is not within the boundary range, the expert experience knowledge needs to be corrected.

The intelligent adaptive control method for hydrogen production by electrolysis of water that is adaptable to wide power fluctuations proposed in the present invention can adapt to the wide power fluctuations of the alkaline electrolyzer caused by the uncertainty of renewable energy power generation. Through the intelligent control method, the given parameters of the electrolyzer (system pressure, alkali solution flow rate, alkali solution temperature) are adaptively controlled to keep the important parameters of the electrolyzer system (such as hydrogen content in oxygen) within a safe range, thereby effectively improving the hydrogen production efficiency and purity of the electrolyzer while ensuring the safe and stable operation of the electrolyzer system.

The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should fall within the protection scope of the claims of the present invention.

Claims (7)

1. Intelligent self-adaptive control system for preparing hydrogen by electrolyzing water and adapting to wide power fluctuation is characterized by comprising:

expert experience knowledge base module: the method comprises the steps of acquiring expected output parameters of the electrolytic tank under the condition that boundary conditions are met, wherein the boundary conditions are used for limiting system parameters and guaranteeing safe operation of the system, and the expected output parameters are used for enabling the parameters during the operation of the system to meet the set boundary conditions;

feedback compensation module: the method comprises the steps of detecting a deviation value between an output parameter of the electrolytic tank after operation and the boundary condition, and outputting an expert experience compensation value;

and (3) a water electrolysis hydrogen production module: the system is used for controlling the stable operation of the electrolytic tank under the condition that the input power fluctuates according to the expert experience compensation value and the output value of the expert experience knowledge base;

the expert experience knowledge base module comprises:

expert experience knowledge base model: for determining, based on the power input by the current electrolytic cell and the current measured operating parameter Y of the electrolytic cell _k Under the limitation of the boundary condition, acquiring an expected output parameter of the electrolytic tank under the limitation meeting the boundary condition; the boundary conditions include: the hydrogen content in oxygen, the oxygen content in hydrogen, the alkali liquor flow, the alkali liquor temperature and the liquid level difference of the oxygen separator and the hydrogen separator; the expected output parameters comprise a system pressure given value, an alkali liquor concentration given value and an alkali liquor flow given value;

the operation measured parameter Y of the electrolytic tank _k Comprising the following steps: system pressure, alkali liquor flow, electrolyzer temperature, liquid level difference of an oxygen separator and a hydrogen separator, content of hydrogen in oxygen in an oxygen scrubber and content of oxygen in hydrogen in the hydrogen scrubber.

2. The intelligent self-adaptive control system for water electrolysis hydrogen production adapting to wide power fluctuation according to claim 1, wherein the expert experience knowledge base module further comprises a query and matching unit, the query and matching unit is used for computing and querying and matching expert experience knowledge in a knowledge base conforming to the boundary conditions through the current operation characteristic parameters of the electrolytic tank based on a multi-attribute similarity algorithm, and screening out expected output values closest to the current working conditions of the electrolytic tank in the knowledge base.

3. The intelligent self-adaptive control system for water electrolysis hydrogen production adapting to wide power fluctuation according to claim 2, wherein the multi-attribute similarity algorithm comprises a nearest neighbor algorithm, euclidean distance and structural similarity, the overall similarity between the current operation parameters of the electrolytic tank and the expert experience knowledge is obtained based on the multi-attribute similarity algorithm, and query and matching are performed based on the similarity.

4. The intelligent self-adaptive control system for the electrolytic water hydrogen production adapting to wide power fluctuation according to claim 1, wherein the expert experience knowledge base module further comprises an evaluation and correction unit and a storage and addition unit, the evaluation and correction unit is used for evaluating and correcting an expert experience knowledge reuse result in the expert experience knowledge base, judging whether to correct the expert experience knowledge according to whether important parameters meet the boundary conditions when the electrolytic tank operates, if the important parameters exceed the boundary conditions, the expert experience knowledge needs to be corrected, and if the important parameters are all within the boundary conditions, the expert experience knowledge does not need to be corrected; wherein the important parameters comprise the content of hydrogen in oxygen, the content of oxygen in hydrogen, the liquid level difference and the alkali liquor temperature; the storage and addition unit is used for adding new expert experience knowledge.

5. The intelligent adaptive control system for producing hydrogen from electrolyzed water adapted to wide power fluctuation according to claim 4, wherein the feedback compensation module comprises:

expert rule establishing unit: the expert gives a correlation coefficient of parameter compensation for each deviation according to experience and rules according to the deviation setting rule between the important parameter and the boundary value which are output after the operation of the electrolytic tank; wherein the rule comprises: an error between the hydrogen content in the oxygen in the operation of the electrolytic cell and the threshold value of the boundary value, an error between the oxygen content in the hydrogen in the operation of the electrolytic cell and the threshold value of the boundary value, an error between the liquid level difference in the operation of the electrolytic cell and the threshold value of the boundary value, and an error between the alkali liquid temperature in the operation of the electrolytic cell and the threshold value of the boundary value;

an inference engine unit: and the expert experience compensation value is deduced by adopting an exhaustive item-by-item search algorithm according to the difference value between the important parameter output in the running of the electrolytic tank and the boundary value through expert experience rules, and the expert experience compensation value is output.

6. The intelligent self-adaptive control system for preparing hydrogen by electrolyzing water, which is adaptive to wide power fluctuation, according to claim 1, wherein the electrolytic water hydrogen preparation module is used for performing transient PID control on a pressure regulating valve and a pneumatic regulating valve in the electrolytic tank through combination of fuzzy rules and a neural network, so as to ensure that the pressure, the liquid levels of an oxygen separator and a hydrogen separator and the alkali liquid temperature and flow rate of the electrolytic tank at the current moment respond to given values tending to important parameters.

7. The intelligent self-adaptive control method for preparing hydrogen by electrolyzing water, which is suitable for wide power fluctuation, is characterized by comprising the following steps:

constructing an expert experience knowledge base model, and obtaining expected output parameters of the electrolytic cell system under the limit of meeting boundary conditions according to the input power and the actual measurement parameters of the current electrolytic cell, wherein the boundary conditions comprise: the expected output parameters comprise a system pressure set value, an alkali liquor temperature set value and an alkali liquor flow set value;

operating the electrolytic tank system based on the expected output parameters, detecting a deviation value between the output parameters of the electrolytic tank system after operation and the boundary conditions through a feedback compensation model, and outputting expert experience compensation values;

and controlling a pressure regulating valve and a pneumatic regulating valve in the electrolytic tank to carry out transient PID control through combination of a fuzzy rule and a neural network according to the expert experience compensation value and the expected output parameter, and simultaneously monitoring and correcting the output boundary important parameter detection value to a feedback compensation model and an expert experience knowledge base model at the same time.

Images