Method for operating an electrolysis plant, control unit

Abstract

The invention relates to a method for operating an electrolysis plant (1) for the production of hydrogen, comprising at least one stack (2) which is supplied with electricity via a power supply (3) and with water via a water circuit (4), wherein the water is used as a reactant and as a coolant. According to the invention, a load change is predicted by means of predictive information and, before the occurrence of the predicted load change, the volume flow supplied to the at least one stack via the water circuit is adjusted to the predicted load change. The invention further relates to a control unit for executing process steps.

Description

[0001] The invention relates to a method for operating an electrolysis plant for producing hydrogen with the features of the preamble of claim 1. Furthermore, the invention relates to a control unit for carrying out method steps.

[0002] The preferred field of application of the invention is the production of green hydrogen, which is produced using renewable energy sources. State of the art

[0003] Electrolysis is a chemical process in which electrical current is used to split a chemical compound. Hydrogen (H2) and oxygen (O) can be extracted from water (H2O).

[0004] Electrolysis is carried out using an electrolysis system. This system comprises at least one stack comprising a multitude of stacked and interconnected electrolysis cells. Examples include PEM electrolysis cells (PEM stands for "Proton Exchange Membrane"), solid oxide electrolysis cells (SOECs), alkaline electrolysis cells, and AEM electrolysis cells (AEM stands for "Anion Exchange Membrane"). In addition to the at least one stack, an electrolysis system comprises a power supply, a water circuit, a gas conditioning system, and a control system with the associated components and lines.

[0005] As hydrogen becomes increasingly important as an energy carrier, the current goal is to produce hydrogen through electrolysis on a large, industrial scale, preferably with plants with a capacity of more than 1 GW. Such plants require multiple stacks or electrolysis modules, with each electrolysis module consisting of a single stack or a complete system, i.e., a stack including power supply, water cycle, gas processing, and control.

[0006] Water is supplied to at least one stack via the water circuit, serving as a reactant and usually also as a coolant. Depending on the operating strategy, a specified minimum amount of coolant is required during partial and full-load operation of at least one stack to ensure that a defined delta temperature is not exceeded across the stack. This temperature limit specification serves to protect against overheating and thus against damage and / or degradation of the stack.

[0007] The volume flow through the stack has a low dynamic compared to the electrical load, so that the time of the load point change can be delayed due to the specified temperature limit.

[0008] The present invention addresses the task of adapting the operating strategy of an electrolysis plant so that load point changes can be implemented with greater dynamism. At the same time, during steady-state operation, only the required volume flow is pumped through the water circuit.

[0009] To achieve this objective, the method is proposed with the features of claim 1. Advantageous further developments of the invention are set forth in the subclaims. Furthermore, a control unit for executing steps of the method is specified. Disclosure of the invention

[0010] A method is proposed for operating an electrolysis plant for producing hydrogen, comprising at least one stack. The stack is supplied with electricity via a power supply and with water via a water circuit, with the water being used as both a reactant and a coolant. According to the invention, a load step is predicted using prediction information, and before the predicted load step occurs, the volume flow supplied to the at least one stack via the water circuit is adjusted to the predicted load step.

[0011] The proposed prediction of a load step allows for early adjustment of the flow rate in the water circuit, and thus through the stack, so that ideally the required coolant quantity is available at the time of the actual load step. This significantly increases the dynamics of the load point change, as it is decoupled from the dynamics of the flow rate adjustment. At the same time, during steady-state operation, no more flow is delivered than the required flow rate.

[0012] In a further development of the invention, it is proposed that the coolant requirement, which changes as a result of the load change in order to maintain a defined delta temperature across the stack, is determined and taken into account when adjusting the volume flow. With knowledge of the coolant requirement, the adjustment of the volume flow can be started in good time, since the duration of the adjustment process can be determined from the coolant requirement. In this way, it can be ensured that the amount of coolant required to maintain the defined delta temperature is available at the time of the actual load change. In this case, the load point change can be made almost without delay. Maintaining the defined delta temperature, in turn, prevents overheating and the associated damage and / or degradation of the stack.

[0013] Furthermore, it is proposed that renewable energy sources, in particular solar and / or wind, be used to supply power to at least one stack. The proposed process can thus produce green hydrogen. However, a disadvantage is that renewable energy sources are not always available and / or are not available in the same quantities. The output of an electrolysis plant powered by renewable energy sources can therefore vary considerably. This means that load changes or sudden load surges occur more frequently.

[0014] As a further development measure, it is therefore proposed that weather data, preferably local weather data, and furthermore preferably local weather data regarding sunshine duration and / or wind speed, be used as prediction information. When using renewable energy sources, changes in energy availability can be derived from this weather data. Depending on the energy availability, the operation of the electrolysis plant can then be adjusted, in particular, a load change can be implemented.

[0015] Furthermore, electricity prices can be used as predictive information. Since electricity prices, especially when using renewable energy sources, depend on the time of day, load changes are implemented to take advantage of times when electricity is particularly cheap. Knowing the electricity prices, especially an upcoming change in the electricity price, the load change can then be predicted.

[0016] Furthermore, it is proposed that the availability of the stacks within a multi-stack system be used as prediction information. This assumes that the electrolysis plant is a multi-stack system that can be operated with a variable number of stacks. For example, if a stack needs to be put into standby mode or if a stack is about to fail, the missing power can be compensated by at least one additional stack, which, however, requires a load change. Knowing the availability of the stacks allows this load change to be predicted, allowing the volume flow in the water circuit to be adjusted in good time.

[0017] Another reason for a load change or load jump can be a changing hydrogen demand requested by a hydrogen-powered unit, such as a fuel cell system or a gas engine. The prediction information can therefore also refer to operating data from a hydrogen-powered unit. The same applies analogously to manufacturing industries where production is ramped up according to plan, such as ammonia synthesis, so that a load jump can be predicted.

[0018] Preferably, the flow rate in the water circuit is adjusted via the flow rate of a pump integrated into the water circuit. For example, the pump speed can be varied to adjust the flow rate to a predicted load step.

[0019] To implement the method, a control unit for an electrolysis plant is further proposed, which is configured to carry out steps of the proposed method. With the aid of the control unit, in particular, a load step can be predicted. For this purpose, prediction information is made available to the control unit from an information source, for example an external server. The control unit preferably has a suitable communication interface for receiving the prediction information. Furthermore, with the aid of the control unit, the volume flow in the water circuit can be adapted to a predicted load step. In particular, a pump integrated into the water circuit can be controlled in order to adapt the pump's flow rate to the predicted load step.

[0020] The invention and its advantages are described in more detail below with reference to the accompanying drawings. These show: Fig. 1 a schematic representation of an electrolysis plant for the production of hydrogen, Fig. 2 a schematic representation of the curves a) of the electrical power P, b) of the coolant volume flow Q and c) of the delta temperature ΔT across the stack, each at a load step without prediction and Fig. 3 a schematic representation of the curves a) of the electrical power P, b) of the coolant volume flow Q and c) of the delta temperature ΔT across the stack, each during a load step with prediction. Detailed description of the drawings

[0021] The Fig. Figure 1 shows an example of an electrolysis plant 1 which is suitable for carrying out the proposed process or can be operated according to the proposed process.

[0022] The electrolysis plant 1 shown comprises at least one stack 2 (only one stack 2 is shown, but the number of stacks 2 can also be two or more), a power supply 3, and a water circuit 4 into which a pump 5 is integrated. Water is supplied to the stack 2 via the water circuit 4 with the aid of the pump 5. The water is split into hydrogen and oxygen in the stack 2 using the electrical energy provided by the power supply 3. The oxygen produced in this way is fed, together with residual water, to a first gas-liquid separator 8. The water separated in the gas-liquid separator 8 is then fed back to the stack 2 via the water circuit 4. The hydrogen produced further is fed to a second gas-liquid separator 9. There, the hydrogen is purified, separating water. This water is also recirculated via the water circuit 4.A cooler 6 and an ion exchanger 7 are integrated into the water circuit 4 downstream of the pump 5. These serve to treat the water. The water supplied to the stack 2 via the water circuit 4 serves not only as a reactant but also as a coolant.

[0023] The power supply 3 of the electrolysis plant 1 shown here is supplied with electrical energy – purely by way of example – via a wind turbine 10. The electrolysis plant 1 shown here therefore produces green hydrogen because it uses a renewable energy source. Other renewable energy sources can also be used. Since renewable energy sources, such as wind, are not always available, the operation of the electrolysis plant 1 must be adapted to the energy availability. This results in load changes or sudden load changes, which require an adjustment of the volume flow in the water circuit 4.

[0024] The Fig. 2 are examples of the curves of the electrical power P ( Fig. 2a)) and the volume flow Q ( Fig. 2b)) in the case of a load jump, which is due to a higher electrical power demand at time t1 (see curve A). In order to achieve a defined maximum delta temperature ΔT max not to exceed the stack 2 (see Fig. 2c)), the volume flow Q in the water circuit 4 must be increased at time t1. However, this requires a certain amount of time (see curve C in the Fig. 2b)). The increase in electrical power P follows the dynamics of the volume flow change (see curve B in the Fig. 2a)), so that the power adjustment occurs with a significant time delay. Using the proposed method, which is described below using the Fig. 3, this can be prevented.

[0025] In the proposed method, the load change at time t1 is predicted using prediction information. This can, in particular, be weather data that indicates when electrical energy can be provided via the wind turbine 10. The operation of the electrolysis plant 1 is then adjusted to the energy availability. Load jumps can be predicted in this way.

[0026] If a load jump is predicted at time t1, the adjustment of the volume flow Q can be started early. Preferably, the coolant demand at time t1 is first determined, from which the time Δt can be calculated, which is needed to provide the required coolant volume flow Q at time t1. If Δt is known, a time t 1' be determined by starting to increase the volume flow Q (see curve C in the Fig. 3b)). By means of prediction, it can be assumed that the increased electrical power P is already present at time t 1' requested (see curve A' in the Fig. 3a)). By adjusting the volume flow Q in advance, the electrical power P can be increased independently of the dynamics of the volume flow change (see curve D in the Fig. 3a)), i.e. almost without delay.

[0027] How Fig. 3c) shows, the prediction not only prevents an exceedance of ΔT max safely prevented, but ΔT max is even temporarily undercut (see curve E in the Fig. 3c)).

[0028] Furthermore, hydrogen production can be increased because the load jump occurs almost immediately. This is exemplified in the Fig. 3d), where the amount of hydrogen gained is indicated by the hatched area between curves D and B.

Claims

[1] Method for operating an electrolysis plant (1) for producing hydrogen, comprising at least one stack (2) which is supplied with electricity via a power supply (3) and with water via a water circuit (4), the water being used as a reactant and as a coolant, characterized by that a load jump is predicted by means of prediction information and before the occurrence of the predicted load jump, the volume flow supplied to the at least one stack (2) via the water circuit (4) is adapted to the predicted load jump. [2] Method according to claim 1, characterized by that by predicting a load jump, the changing coolant requirement caused by the load jump is determined to maintain a defined delta temperature across the stack and is taken into account when adjusting the volume flow. [3] Method according to claim 1 or 2, characterized bythat renewable energy sources, in particular solar and / or wind, are used for the electricity supply (3). [4] Method according to one of the preceding claims, characterized by that weather data, preferably local weather data, further preferably local weather data regarding sunshine duration and / or wind strength, are used as prediction information. [5] Method according to one of the preceding claims, characterized by that electricity prices are used as predictive information. [6] Method according to one of the preceding claims, characterized by that the availabilities of the stacks (2) within a multi-stack system are used as prediction information. [7] Method according to one of the preceding claims, characterized by that the volume flow in the water circuit (4) is adjusted via the flow rate of a pump (5) integrated into the water circuit (4). [8] Method according to one of the preceding claims, characterized by that the operating data of a hydrogen-powered unit, for example a fuel cell system, are used as prediction information. [9] Control unit for an electrolysis plant (1), wherein the control unit is configured to carry out steps of a method according to one of the preceding claims.

Images