Field of the invention

The invention relates to operating a heat releasing reactor.

Technical background

A commonly used technical field of heat releasing reactors is combustion boilers, such as grate boilers and fluidized bed boilers are commonly utilized to generate steam which can be used for variety of purposes, such as for producing electricity and heat.

In a fluidized bed boiler, fuel and solid particulate bed material is introduced into a furnace. The bed material and fuel is fluidized by introducing fluidizing gas from a bottom portion of the furnace. Burning of fuel takes place in the furnace. In BFB combustion, fluidization gas is passed through the bed such that it forms bubbles in the bed. The fluidized bed can in a BFB be rather conveniently controlled by controlling the fluidization gas feed and fuel feed. In addition to fuel, certain additives such as aluminum silicates (such as, non-hydrated clay) and alkali alkaline earth metal carbonates and mixtures thereof (such as, limestone or calcium carbonate) may be added to the combustion to improve sorption of possible heavy metals, sulfur, and also to improve alkali sorption.

In CFB combustion, fluidization gas is passed through the bed material. Most bed particles will be entrained in the fluidization gas and they will be carried with flue gas. The particles are separated from the flue gas in at least one particle separator and circulated returning them back into the furnace. It is common to arrange a fluidized bed heat exchanger downstream the particle separator(s) to recover heat from the particles before they are returned into the furnace.

In all boilers, regardless of the combustion technology, the combustion conditions, such as, the mixing of air and fuel, may not be ideal.

International application published under WO 2016/202640 Al of Improbed AB discloses a thermal load control method for a combustion boiler. In the method, the thermal load of a combustion boiler is reduced if monitored flue gas velocity in at least one location of the boiler exceeds a pre-determined maximum flue gas velocity limit. The flue gas velocity is computed from volume flow of flue gas divided by the cross- sectional area of the flue gas duct in the location just downstream the cyclone using an equation group.

Additionally, it is known to there are also other processes, which produce product gas, temperature of which needs to be controlled, meaning either heating or cooling the gas and/or the process.

Objective of the invention

A heat releasing reactor is designed for a given capacity that is the respective maximum continuous rating (MCR) of the reactor. This is sometimes called the design load level.

It is a particular objective of the invention to improve performance, profitability, and flexibility of the heat releasing reactor, and also to improve control of the reactor's load. This objective can be achieved with the method of operating a heat releasing reactor according to claim 1 and with the heat releasing reactor according to claim 19.

A further objective of the invention is to reduce complexity of control system of a combustion boiler. This objective can be met with a reactor computation system according to claim 24. The dependent claims describe advantageous aspects of the method, of the reactor, and of the computation system.

Advantages of the invention

The method of operating a heat releasing reactor producing product gas comprises the steps of a) monitoring the current load Q _h of the reactor; b) finding such a numerical value for a current computational maximum momentary load for which at least one product gas factor computed using currently monitored process data with a numerical model of the reactor fulfills an acceptance condition, and selecting the numerical value as the current computational maximum momentary load Qh,max; c) indicating the current computational maximum momentary load Qh,max to the operator and/or, if the current load Q _h is cl) smaller than the current computational maximum momentary load: cli) indicating the operator that the load may be increased, and/or clii) automatically increasing the load, and/or c2) larger than the current computational maximum momentary load: c2i) indicating the operator that the load Q _h exceeds the current computational maximum momentary load, and/or c2ii) automatically reducing the load Qh.

With the method, instead of having a fixed maximum load, with the method of computing the product gas factor and selecting its acceptance conditions suitably, it is possible to safely operate the reactor at or closer to its current computational maximum momentary load that at times may be higher than the fixed maximum load would be. The current computational maximum momentary load can be higher than the design load level. Therefore, the overall performance of the reactor may be improved enabling increased power/heat production. Further, since the current computational maximum momentary load may occasionally be smaller than the design load level, wear of the reactor resulting from exceeding the current computational maximum momentary load may be better reduced. In other terms, the current computational maximum momentary load can be considered as maximum allowable load and/or preferable load.

The applicant has been able to obtain, in the tests performed with a boiler, in average, power output from a combustion boiler that exceeds the fixed boiler maximum load. The applicant could in the tests demonstrate that for a combustion boiler the improvement potential may be between 2,5 - 5% which corresponds, for example, 3 to 6 MWth for a 120 MWth combustion boiler.

Preferably, in the method: i) the currently monitored process data of the reactor includes ia) current product gas exit temperature in a product gas flow channel and ib) heat duty for each heat transfer surface in the product gas flow channel and further: ii) monitored process data from both steps ia) and ib) is used in computation of the product gas factor and when finding the numerical value for the current computational maximum momentary load Qh,max. Computation of heat duty of a heat exchanger is known for skilled person in the art and heat duty can be obtained, for instance, by using the following equation

Q _fluid,i = q _m,fluid,i *(h _fluid,out - h _fluid,in )' wherein q _mi fi _U id,i is the fluid flow in ith heat transfer surface, h _fiuid,in is the enthalpy of fluid entering to the ith heat transfer surface and h _fiuid,out is the enthalpy of fluid exiting from the ith heat transfer surface.

The finding may be performed such that, if the at least one product gas factor computed using currently monitored process data with a numerical model of the reactor fails to fulfill an acceptance condition, a next numerical value is automatically selected. Preferably, the next numerical value is selected iteratively. This may enable the use of computational library functions, and, in particular of an iterative solver (such as, Python FSOLVE function which solves roots of function).

The finding may be carried out with performing the computational steps of:

- I: computing an estimate for boiler product gas exit temperature that results in a computational reactor model when the thermal load of the reactor corresponds to the numerical value;

- II: computing product gas mass flow

- Ill: computing a heat duty for each heat transfer surface in the flue gas flow channel using its current heat duty that is corrected by using a numerical boiler model;

- IV: using the computed heat duties for each heat transfer surface in the product gas flow channel to compute product gas temperatures at each heat transfer surface in the product gas flow channel in the upstream direction of gas flow, starting from the heat transfer surface that is closest to the product gas exit using the estimate for the flue gas exit temperature; - V: computing a product gas factor for each heat transfer surface in the flue gas flow channel. With this approach, the situation of each heat transfer surface (in this application “heat transfer surface” means a heat exchanger, a heat exchanger tube, heat exchanger tube bundle, heat exchanger packages and/or a constructive group of heat exchangers) in the product gas flow channel can be estimated numerically with the product gas factor in the situation where the thermal load of the reactor corresponds to the numerical value. So, we can now test whether a given numerical value that is a candidate for a current computational maximum momentary load would produce an acceptable situation at the heat transfer surface. According to an embodiment of the invention, in step III) the numerical reactor model is of the form Q _{fluid, i, candidate} = The fitting of the parameters (α _j,i )can be done manually by human or automatically by computer utilizing historical data. Automatic update of the parameters may be done e.g. once per month. AI and neural network based algorithms can be utilized in automatic update. Particularly when the invention is applied to a boiler combusting fuel, this enables predicting the maximum computational allowable current boiler momentary load without going to the limit with the current boiler load, in contrast to the method disclosed in WO 2016/202640 A1, and on the other hand, and even more importantly, enables going to the limit without exceeding the maximum computational allowable current boiler momentary load. Preferably, the p d f i l d i where k _i is a non-zero parameter that may be chosen combustion- boiler specifically, preferably positive (non-zero) number q _{m,product gas} is a product gas mass flow n is a model parameter that may be chosen reactor- specifically, being preferably positive (non-zero) number ρ _productgas,i is the density of the product gas at the i ^th heat transfer surface; and A _cross,i is the cross-sectional area of the gas flow path at the i ^th heat transfer surface. This is particularly convenient since choosing this functional form for the product gas factor, it becomes very flexible and can be easily adapted to suit different needs of operation of the reactor, such as, based on the conditions in the current reactant used in the reactor. i) Particularly advantageously, the model parameter n may be selected to include at least one of the following: in the range of 0,9 to 1,1, preferably about 1.0, for using computed product gas velocity; ii) in the range of 2,9 to 3,5, preferably between 3,2 and 3,35, for using computed product gas caused erosion; or iii) in the range of 1,8 to 2,2, preferably about 2.0, for using pressure loss of the product gas flow. The value for n may be changed over time. This is advantageous for the reason that the flue gas flow conditions at the heat transfer surfaces may change over time, such as because of fouling, ash agglomeration or reactant or bed conditions, should the reactor be a fluidized bed reactor. Thus, the product gas factor may be shifted over time, to better reflect the actual process situation.

According to an embodiment of the invention, when n=2 and product gas factor represents a pressure loss, the comparison between the product gas factor dfi and a predetermined maximum value for the product gas factor df _max,i can be carried out for each heat transfer surface. According to an embodiment the acceptance condition is substantially

According to an embodiment of the invention, when n=2 and product gas factor represents a pressure loss, the comparison can be done between the sum of the flue gas factors d _fi and sum of the predetermined product gas factors df _max,i or simply predetermined product gas factor represents total pressure drop and hence the comparison represents the comparison of total pressure drops between the reactor and stack. According to an embodiment the acceptance condition is substantially dp _tot = dp _max,tot.

According to an embodiment of the invention, the product gas factor represents an particle loading factor and can be written in the form where k _Ph is particle hardness factor, C(d) is particle diameter function, q _m _ _fa is particle mass flow rate, v _p is particle velocity and n is exponent (0,3 - 4). In such case, the predetermined product gas factor represents maximum particle loading value. It can also be adjustable based on the particle properties (softness, etc.).

According to a specific embodiment of the invention, the reactor is a fluidized bed (FB) reactor, such as a FB boiler or FB gasifier, the product gas factor represents an ash loading factor and can be written in the form where k _Ph is particle hardness factor, C(d) is particle diameter function, q _m-fa is fly ash mass flow rate, v _p is particle velocity and n is exponent (0,3 - 4). In such case, the predetermined flue gas factor represents maximum ash loading value. It can also be adjustable based on the ash properties (softness, etc.).

According to an embodiment of the invention, the acceptance condition is substantially df _i = df _max,i but in practical circumstances the acceptance condition can be defined as wherein 5>0 and depends on the numerical accuracy and/or method. When it means that at least one product gas factor computed using currently monitored process data with a numerical model of the reactor fulfills the acceptance condition and in such a case maximum allowable load has been found and so the numerical value Qh,candidate is selected as the current computational maximum momentary load Qh,max•

According to an embodiment of the invention, the acceptance condition is substantially Σ (dfi)= Σ (dfmax,i) but in practical circumstances the acceptance condition can be defined as utilizing the following sums wherein δ>0 and depends on the numeric accuracy and/or method. When it means that at least one product gas factor computed using currently monitored process data with a numerical model of the boiler fulfills the acceptance condition and in such a case maximum allowable load has been found and so the numerical value Qh,candidate is selected as the current computational maximum momentary load Qh,max• According to an embodiment the summation index i goes over all of the heat transfer surfaces. According to another aspect of the invention, the summation index i goes over only a part of the heat transfer surfaces, preferably in a product gas channel.

It may be particularly useful if the value for n is determined from a group of reactors comprising at least two separate reactors using operational data monitored for each of the reactor. Using a larger number of reactors (two, three, four, ...) gives a larger data set. Hence, there will be more operational data monitored. This may produce better results, which may be especially good in a situation where the determination uses interpolation and/or extrapolation of experimental data.

For the computation in step I), the product gas exit temperature may be substantially estimated by equation or preferably its first, second, or third (or higher) degree approximation. The coefficients α may be obtained by fitting after measuring product gas exit values for a number of discrete reactor load values. This data may be collected over time and refreshed from time to time, such as, periodically. Alternatively, or in addition, it may be collected in one or more calibration runs of the reactor.

The fitting of the coefficients (a) can be done manually by human or automatically by computer utilizing historical data. Automatic update of the coefficients may be done e.g. once per month. Al and neural network based algorithms can be utilized in automatic update.

According to an embodiment of the invention, in step I), the product gas exit temperature may be substantially estimated by utilizing artificial intelligence tools. According to another embodiment of the invention, in step I), the product gas exit temperature may be substantially estimated by utilizing neural network. According to an embodiment of the invention, in step I), the product gas exit temperature may be estimated by equation wherein α0, α1 and α2 can be predefined constants. Alternatively or in addition, the fitting of the coefficients (a) can be done manually by human or automatically by computer utilizing historical data. Automatic update of the coefficients may be done e.g. once per month. Al and neural network based algorithms can be utilized in automatic update.

According to an embodiment of the invention, oq term may be solved based on the current state values wherein T _{G,exit current} represents measured product gas exit temperature.

According to an embodiment of the invention, in step II), the product gas mass flow is computed using reactor mass and energy balance equations.

In step II), the computation of product gas mass flow may include taking into account gas-specific mass flow of components of the product gas. In case of combustion process the components comprise CO ₂ , H ₂ O, N ₂ , SO ₂ , O ₂ . The concentration of these components can be measured reliably with rather simple equipment.

In step II), the component values may include reactant parameters. This enables reflecting changes in the reactant properties. For example, for reactants that tend to cause more erosion, the acceptance condition may be stricter, while leaner acceptance condition may be used for reactants that tend to cause less erosion.

The step b) may be performed remotely to the reactor, preferably in a cloud-based computation service. This helps to simplify the maintenance of the combustion boiler, since the remote computation equipment, such as configured to run the cloud-based computation service, can be maintained separately from the combustion boiler. The computational software updates, for example, can thus be performed centrally at one or a few locations, instead of updating software at each reactor.

Alternatively, the step b) may be performed locally at the site of the reactor, preferably by an edge server. This may speed up the computation since no data needs to be transferred to a remote computation location.

Any of the currently monitored process data and/or current load may be obtained from real-time measurements. Instead of this, or in addition to it, the currently monitored process data and/or current load may be treated by filtering, treated by averaging, computing trends or any combination of these. This helps to avoid noise or outlier measurements to impact the outcome of the computation, and thus facilitates to increase stability of the current computational maximum momentary load.

The acceptance condition may include a hysteresis condition, requiring a predefined minimum change before changing the current computational maximum boiler momentary load. This may increase the stability of the current computational maximum momentary load, preferably helping to avoid excessive fluctuation of the current computational maximum momentary load.

The applicant finds it also useful if the combustion boiler is a circulating fluidized bed (CFB) or a bubbling fluidized bed (BFB) boiler, as well as more generally CFB and BFB reactors and the step b) is carried out for the fluidized bed reactor heat transfer surfaces.

According to an embodiment the step b) is carried out for the heat transfer surfaces between a reactor and a stack.

A heat releasing reactor comprises: - a reaction chamber and associated passes defining a product gas flow path and having a number of heat transfer surfaces;

- measurement instrumentation to monitor current load of the heat releasing reactor;

- further measurement instrumentation to currently monitor process data; and

- a control system configured to carry out the method of operating the heat releasing reactor.

According to an embodiment heat releasing reactor comprises a reactor chamber and associated passes defining a gas flow path and having a number of heat transfer surfaces in the gas flow path.

Such a heat releasing reactor can improve the control aspect of the reactor. The advantages are same as the advantages of the method of operating the heat releasing reactor.

The control system may comprise an edge server which may be configured to process the real-time measurement results for currently monitored process data and/or current load, namely by filtering, averaging, and/or computing trends. The edge server will facilitate cutting down the amount of currently monitored process data. In certain installations this may be particularly useful especially in view of the fact that there may be 60 to 90 gigabytes of monitored process data each day.

The control system may be configured to carry out the method step b) to determine the current computational maximum momentary load locally. This facilitates to have fast decision making at the reactor since less or no data may need to be transferred from the reactor.

Alternatively, or in addition, the control system may be configured to send data to a remote, preferably cloud-based, computing system which may be configured to carry out the method step b) and return the current computational maximum boiler momentary load to the control system. This facilitates to have a reactor simpler and makes updating the computing system easier. The updating can in this situation be performed centrally and not at each and every reactor.

The edge server may be configured to reduce amount of measurement data that is passed to the remote computing system. In this manner, a smaller bandwidth for transferring data may suffice. In certain installations this may be particularly useful especially in view of the fact that there may be 60 to 90 gigabytes of monitored process data each day.

A reactor computation system comprises

- a group of reactor, each reactor comprising a reactor control system comprising an edge server system which is configured to process the real-time measurement results for currently monitored process data and/or current load, by filtering, averaging, and/or computing trends, and send the processed real-time measurement results to a remote computing system;

- a remote computing system which preferably is a cloud- based computing system, configured to receive data processed from real time measurement results and to compute data using a numerical reactor model for each of the reactors, and return computation results for each of the reactors.

Further, in the reactor computation system, the control system is configured to adapt its function based on the computation results.

The advantage for this arrangement is that the need of computation devices at the reactor can be reduced, still obtaining effective and fast computation results from the remote computing system.

The computing system may be configured to find such a numerical value or a current computational maximum momentary load for which at least one product gas factor computed using currently monitored process data with a numerical model of the reactor that fulfills an acceptance condition and selecting the numerical value as the current computational maximum momentary load. This basically enables using the method of the invention also in a distributed environment.

The reactor computation system may be configured to calibrate a numerical model, such as, the product gas factor numerical model, for a reactor using processed measurement data for the reactor. This makes it easier to remotely adapt or calibrate the numerical model for reactor control.

The reactor computation system may be configured to adapt or calibrate a numerical model for a reactor using processed measurement data collected also from other reactors. This enables using a larger collection of data to adjust the numerical model for reactor control.

The invention is applicable to be used in various processes and reactors, where heat generation and its recovery is involved.

Invention is applicable to be used in connection with thermochemical reactor for controlling its load, that is a rate of charging or discharging energy from the reactor.

Invention is further applicable to be used in connection with a waste gasifier for controlling the heat produced during the thermochemical destruction of waste while producing gas. Such gas can be further synthesized into fuels and chemicals in a following synthesis process.

Invention is further applicable to be used in connection with a so-called carbon capture process so as to control the load of a looping carbon capture reactor (e.g. calcium looping reactor) for additional heat production while capturing CO ₂ from a gas flow. List of drawings

The reactor and its control method are explained in more detail below in the context of the embodiments shown in the appended drawings in FIG 1 to 9 , of which :

FIG 1 illustrates a CFB boiler ;

FIG 2 illustrates a BFB boiler ;

FIG 3 illustrates the flow of measurement data from sensors ;

FIG 4 is a flow diagram illustrating a first method for finding the current computational maximum boiler momentary load Q _{h, mx} ;

FIG 5 is a flow diagram illustrating a second method for finding the current computational maximum boiler momentary load Q _{h, max} ;

FIG 6 illustrates how the current computational maximum boiler momentary load Qh, max can be presented to the boiler operator ;

FIG 7 shows boiler momentary load Q _h and computed current computational maximum boiler momentary load Qh, max , as well as the effect of using the method according to the invention during a test period;

FIG 8 a closer look at the data of FIG 7 , showing boiler momentary load Q _h computed current computational maximum boiler momentary load Q _{h, max} where the effect of us ing the method according to the invention during the 10 day test period is better visible ;

FIG 9 illustrates another heat releasing reactor according to an embodiment of the invention .

Same reference numerals refer to same technical features in all figures . Detailed description

FIG 1 shows a combustion boiler 10 operating as a heat releasing reactor producing heat in a form of a circulating fluidized bed reactor CFB. CFB reactor may be used as a combustor/steam generator, a carbon capture reactor using (calciner and/or carbonator), as well as a waste material gasifier. In the following description the CFB reactor is referred particularly as a circulating fluidized bed (CFB) boiler and comprises a furnace 12 that has tube walls 13 connected to water-steam circuit of the combustion boiler 10. Water is fed from water tank (not shown) to economizer and from the economizer via a steam drum to evaporative heat transfer surfaces such as the tube walls 13 and then guided via the steam drum to superheaters and then to a turbine. Flue gas channel may be provided with economizer and/or superheater/s.

Fluidization gas (such as, air and/or oxygen-containing gas) is fed from fluidization gas supply 153 to below the grate (the grate not shown in FIG 1) via a windbox (not shown), wherefrom the primary fluidization air enters into the furnace through nozzles (not shown) (to fluidize the bed), and secondary fluidization gas feed 152 (to feed oxygen containing gas to control combustion). The effect is that the bed materials will be fluidized and also oxygen required for the combustion is provided into the furnace 12. Further, fuel is fed into the furnace 12 via the fuel feed 22. The combustion can be adjusted by controlling the fuel feed 22 (such as, by reducing or increasing fuel feed), and by controlling the fluidization gas feed (such as, by reducing or increasing amount of oxygen supply into the furnace 12). Fuel can be fed together with additives, in particular with such additives that act as alkali sorbents, such as CaCO ₃ and/or clay for example. In addition or alternatively, NOx reduction agents, such as ammonium or urea can be fed into the combustion zone of the furnace 12, or above the combustion zone of the furnace 12. Bed material is also fed into the furnace, which bed material may comprise sand, limestone, and/or clay, that in particular may comprise kaolin. One effect of the bed and, generally, of the combustion, is that in the water-steam circuit, water and steam is heated in the tube walls 13 and water is converted to steam.

Ash may fall to the bottom of the furnace 12 and be removed via an ash chute (omitted from FIG 1 for the sake of clarity) and part of the ash, so-called fly ash, will be carried along flue gas.

Combustion products, such as flue gas, unburnt fuel and bed material proceed from the furnace 12 to a particle separator 17 that may comprise a vortex finder 103. The particle separator 17 separates flue gases from solids. Especially in larger combustion boilers 10, there may be more than one (two, three, ...) separators 17 preferably arranged in parallel to each other.

Solids separated by the separator 17 pass through a loop seal 160 that preferably is located at the bottom of the separator 17. Then the solids pass to fluidized bed heat exchanger (FBHE) 100 that is also a heat transfer surface so that the FBHE 100 collects heat from the solids to further heat the steam in the water-steam circuit. The chamber in which the FBHE 100 is located may be fluidized and the FBHE 100 itself comprises heat transfer tubes or other kinds of heat transfer surfaces. FBHE 100 may be arranged as a reheater or as a superheater. From the FBHE outlet 101, steam is passed into a high-pressure turbine (if the FBHE 100 is superheater) or medium-pressure turbine (if the FBHE 100 is a reheater). For the sake of clarity, the turbines are not illustrated in FIG 1. The solids may be returned from the FBHE 100 via a return channel 102 into the furnace 12. Especially in larger combustion boilers 10, there may be more than one (two, three, ...) loop seals 160 and FBHE 100, and return channels 102, preferably arranged in parallel to each other, such that for each separator 17, there will be respective loop seal 160, FBHE 100 and return channel 102. In practice, some of the FBHE 100 may be arranged as superheaters while some others may be arranged as reheaters. The flue gases are passed from the separator 17 to horizontal pass 15 and from there further to backpass 16 (that preferably may be a vertical pass) and from there via flue gas conduit 18 to stack 19. The backpass 16 comprises a number of heat transfer surfaces 21 _i (where i = 1, 2, 3, …, k, where k is the number of heat transfer surfaces). In FIG 1, heat transfer surfaces 21 ₁ , 21 ₂ , 21 ₃ , ..., 21 _k-1 , 21 _k are illustrated. Heat transfer surface 21 _k depicts air preheater. Heat transfer surfaces 21 _k-1 , 21 ₂ depict superheaters and heat transfer surfaces 21 ₁ , 21 ₃ depict reheaters. The actual number of different heat transfer surfaces in each of these components, for example, may be selected for each combustion boiler differently according to actual needs. And there may be further components as well, comprising a heat transfer surface 21. Flue gas exiting the last heat transfer surface 21 _k will be in flue gas exit temperature T _{G, exit} . This temperature is measured with temperature sensor 20 _k . According to one aspect, the temperatures before and after each heat transfer surface 21 _i (T _G,in,i , T _G,in,i+1 , respectively) can be measured with respective temperature sensors 20 _i (where i = 1, 2, 3, …, k-1, k). According to another aspect, and preferably, these temperatures however do not necessarily need to be measured. It will suffice to know the flue gas exit temperature T _{G, exit} . The temperatures before and after each preceding heat transfer surface 21 _i (T _G,in,i , T _G,in,i+1 ) can be obtained numerically. This will be explained further below. A combustion boiler 10 is equipped with a plurality of sensors and computer units. Actually, one middle-size (100 – 150 MW _th ) combustion boiler 10 may produce 100 million measurement results / day, which needs 25 GB of storage space. FIG 1, 2 and 3 illustrate some of the sensors and computer units. Examples of sensors are combustion gas (usually combustion air) volume flow sensors 30 (for measuring primary and secondary fluidizing gas feeds), fuel feed sensors 650 and temperature sensors 20i (i = 1, 2, k), temperature sensor in FBHE and pressure sensor

116 in the return channel 102 (both only in a CEB boiler), and sensors 40 in the furnace 12.

Process data may be collected from the sensors by distributed control system (DCS) 201. The data collection may most conveniently be arranged over a field bus 290, for example. DCS 201 may have a display/monitor 202 for displaying operational status information to the operator. An EDGE server 203 may process measurement data from the obtained from sensors, such as, filter and smooth it. There may be a local storage 204 for storing data.

The DCS 201, display/monitor 202, EDGE server 203, local storage 204 may be in combustion boiler network 280 (local storage 204 preferably directly connected to the EDGE server). The combustion boiler network 280 is preferably separate from the field bus 290 that is used to communicate measurement results from the sensors to the DCS 201 and/or the EDGE server 203. Between the DCS 201 and EDGE server 203 there may be an open platform communications server 210 (cf. FIG 3) to make the systems better interoperable.

Combustion boiler network 280 may be in connection with the internet 200, preferably via a gateway 290. In this situation, measurement results may be transferred from the combustion boiler network 280 to a cloud service, such as process intelligence system 205 located in a computation cloud 206. The applicant currently operates a cloud service running an analysis platform. The cloud service may be operated on a virtualized server environment, such as on Microsoft® Azure® which is a virtualized, easily scalable environment for distributed computing and cloud storage for data. Other cloud computing services may be suitable for running the analysis platform too. Further, instead of a cloud computing service, or in addition thereto, a local or remote server can be used for running the analysis platform.

FIG 2 illustrates a heat releasing reactor which is a combustion boiler 10 that is a bubbling fluidized bed (BFB) boiler. BFB boiler differs from CFB boiler in that the fluidized bed is not a circulating bed but a bubbling bed. Thus, there may not be a need for the separator 17, loop seal 160, FBHE 100 and return channel 102.

There is normally at least one superheater 14 located in the furnace 12, preferably on top of the furnace 12. Superheater 14 inlet 141 is preferably the steam drum or from another superheater and the outlet 142 is to high pressure turbine.

FIG 4 illustrates the method of operating a heat releasing reactor producing product gas: a) the current load Qh of reactor 10 (such a combustion boiler, gasification reactor, liguid air energy storage, hydration reactor) is monitored in step KI (in the method illustrated in FIG 4, also product gas exit temperature T _{G, exit} is monitored and heat duty Q _fluid,i to heat transfer fluid for each heat transfer surface 21i in the gas flow channel 16. b) a numerical value Q _h , is selected (step K3), after which heat duties at heat transfer surfaces 21i are computed and gas temperatures in relation to Q _h,candidate The numerical value Q _h,candidate is then used to compute (step K7) at least one product gas factor dfi (may be referred to as flue gas factor in connection with combustion process)using currently monitored process data with a numerical model of the reactor fulfills an acceptance condition (which is tested in step K9), and selecting the numerical value Q _h,candidate as the current computational maximum momentary load Qh, of the reactor 10(step Kll); c) the current computational maximum momentary load Qh, max is indicated to the operator (such as, by displaying on the monitor/screen 202) and/or, if the current load Q _h is cl) smaller than the computational maximum momentary load Q _h,max : cli) indicating the operator that the boiler load Q _h may be increased, and/or clii) automatically increasing the load Q _h of the reactor, and/or c2) larger than the computational maximum momentary load Q _h,max : c2i) indicating the operator that the load Q _h exceeds the maximum momentary load, and/or c2ii) automatically reducing the reactor load Qh.

In the method, the currently monitored process data of the reactor may include a) current product gas exit temperature T _{G, exit} in a gas flow channel and b) heat duty Qfluid,i for each heat transfer surface 21i in the gas flow channel 16.

Further, in the method monitored process data from both a) and b) may be used in computation of the product gas factor dfi and when finding the numerical value Qh,candidatefor the current computational maximum momentary load Qh,max.

The finding is performed such that, if the at least one product gas factor df _i computed using currently monitored process data with a numerical model of the reactor that fails to fulfill an acceptance condition, a next numerical value Qh,candidate is automatically selected. The automatic selection is preferably done iteratively. As a specific example, the finding may be carried out with performing the computational steps of: - I: computing an estimate for product gas exit temperature T _{G, exit} that results in a computational model when the thermal load of the reactor corresponds to the numerical value Q _{h, candidate;} - II: computing gas mass flow q _m,fluegas;; - III: computing a heat duty Q _{fluid, i, candidate} for each heat transfer surface 21 _i in the flue gas flow channel (back pass 16) using its current heat duty Q _{fluid, i, current} that is corrected by using a numerical boiler model Q _{fluid, i, candidate} = Q _{fluid,i,current} + Σα _j,i (Q _h,candidate ) ^j - Σα _{j,i (} Q _h,current ) ^j - IV: using the computed heat duties Q _{fluid, i, candidate} for each heat transfer surface 21 _i in the gas flow channel 16 to compute gas temperatures at each heat transfer surface (T _G,in,i , T _G,out,i ; i = 1, ... , k) in the gas flow channel 16 in the upstream direction of gas flow, starting from the heat transfer surface 21 _k that is closest to the flue gas exit i.e. using the estimate for the gas exit temperature T _G,out,m = T _{G, exit;} - V: computing a product gas factor df _i , i = 1 , ..., k for each heat transfer surface 21 _i in the flue gas flow channel (back pass 16). The fitting of the parameters (α _j,i )can be done manually by human or automatically by computer utilizing historical data. Automatic update of the parameters may be done e.g. once per month. AI and neural network-based algorithms can be utilized in automatic update. Step II) may include computing product gas mass flow q _m,G,m for selected flue gas components. The gas temperatures at each heat transfer surface can be computed, for instance, wherein T _Gini is the flue gas temperature at the inlet of ith heat transfer surface, c _p is specific heat capacity, and T _Gouti is the flue gas temperature at the outlet of ith heat transfer surface. The flue gas temperatures could be determined with artificial intelligence tools. The flue gas temperatures could be determined with neural network.

Preferably, the product gas factor df _i includes or is: where k _i is a predetermined non-zero parameter that may be chosen combustion-boiler specifically, preferably positive (non-zero) number, qm,G is a flue gas mass flow, n is a positive number (which may be selected as a natural number, rational number, real number, or even as complex number), pG,i is flue gas density obtainable from flue gas temperature TG,in,i at i ^th heat transfer surface 21i and

A is a cross section of flue gas channel at i ^th heat transfer surface 21 _i . i) Advantageously, n may be selected to include at least one of the following: in the range of 0,9 to 1,1, preferably equivalent or about 1.0, for using computed gas velocity; ii) in the range of 2,9 to 3,5, preferably between 3,2 and 3,35, for using computed gas caused erosion; or iii) in the range of 1,8 to preferably equivalent or about 2.0, for using pressure loss. The value for n may be changed over time. In particular, the value for n may be determined from a group of reactors, the group comprising at least two separate reactors 10, such that using operational data monitored for each of the reactors 10 is used in the determination.

In the computation in step I), the computational value for gas exit temperature T _G , under any chosen numerical value Qh,candidate for boiler load can be estimated by equation or preferably its first, second, third or higher degree approximation. The coefficients α ₀ ,α ₁ ,α ₂ ,••• have been obtained beforehand by fitting after measuring flue gas exit temperature T _G , values for a number of discrete reactor load Qh values.

In step II), the computation of the components qm,G,m in case the process in the reactor is combustion of fuel preferably includes at least some, most preferably all of the following: m = CO ₂ , H ₂ O, N ₂ , SO ₂ , O ₂ so as to determine gas mass flow. In other words, in step IV) of the computation, as q _m,G,m values some or all of q _m ,G,CO2, q _m ,G,H20 , q _m ,G,N2 ,qm,G,so2 ,q _m ,G,O2 may be used. They are preferably measured in gas conduit 18 or in stack 19, for which reason suitable sensors are installed in the gas passage. In step II), the component values may further include reactant (such as fuel of alkali oxide, like CaO) parameters.

In case of the reactor is for combustion process the product gas i.e. flue gas mass flow may be based on computation of sums of flue gas component mass flows q _m ,G, _m which are calculated based on fuel analysis (proximate and ultimate analysis of fuel), combustion air flow and/or recirculation gas flow according to boiler mass and energy balance calculation.

Preferably, the flue gas mass flow may be computed: i.e., for example, the sums of the following flue gas mass flow components CO ₂ , H ₂ O, N ₂ , SO ₂ and O ₂ :

where, for instance, x _c,fuel represents carbon in fuel i.e. first subscript denotes component and second subscript is either fuel or combustion air referred, q _m,fuel is a fuel flow,q _m,air is combustion air flow and M _x denotes molar mass. Advantageously, fuel properties as utilized in flue gas mass flow components and combustion air properties. Fuel moisture may be measured or calculated.

According to another embodiment of the invention, when the reactor is a thermochemical reactor, particularly based on CaO /Ca(OH) ₂ hydration / dehydration reaction, the flow of H ₂ O (steam) and air as fluidization gas are the mass flows in the step II), the computation mass flows of which the components are computed making use of hydration and dehydration reactions.

According to another embodiment of the invention, when the reactor is a waste material gasifier the gas flow determination may be provided in a same manner as in case of the combustion process, but the gas composition may be different including also CO and H ₂ and some minor gasification products.

According to another embodiment of the invention, when the reactor is a so called carbon capture reactor, including a fluidized be carbonator and calciner connected with each other suitably. The reactor configured to reduce particularly CO ₂ from a gas to be cleaned (carbonator)by reaction with CaO and producing substantially pure CO ₂ (calciner) by calcination of CaCO ₃ . The gas flow determination may be provided in a same manner as in case of the combustion process, considering that fuel is being burned in a calciner, including gas flow of oxygen and the gas to be cleaned.

The step b) may be performed remotely to the combustion boiler, such as, in the process intelligence system 205. Alternatively, the step b) may be performed locally at the combustion boiler, preferably at the EDGE server 203.

Any of the currently monitored process data and/or current load may be obtained from real-time measurements, treated by filtering, treated by averaging, computing trends or any combination of these.

The acceptance condition may include a hysteresis condition, requiring a predefined minimum change before changing the current computational maximum boiler momentary load

The acceptance condition preferably includes comparing the computed at least one flue gas factor df ₂ against a respective maximum value df _max,2 . The maximum value df _max,2 is a preset value and preferably boiler specific. The numerical value Qh, is rejected if the maximum value df _max,2 is exceeded.

In the combustion boiler 10, the furnace 12 and associated passes (horizontal pass 15 and back pass 16) define a flue gas flow path. The furnace 12 and the passes 15, 16 have a number of heat transfer surfaces 21 _i in the flue gas flow path. The combustion boiler 10 also has measurement instrumentation to monitor current load Q _h of the combustion boiler, and further measurement instrumentation to currently monitor process data.

The control system (DCS 201, and EDGE server 203, or DCS 201 remote process intelligence system 205, possibly under the participation of the EDGE server 203) is configured to carry out the boiler control method.

The EDGE server 203 may be configured to process the real-time measurement results for currently monitored process data and/or current load, namely by filtering, averaging, and/or computing trends.

The control system may be configured to carry out the method step b) to determine the current computational maximum boiler momentary load Q _h,max locally at the combustion boiler 10, and/or to send data to a remote, preferably cloud-based (such as, computation cloud 206), computing system (such as, process intelligence system 205) which is configured to carry out the method step b) and return the current computational maximum boiler momentary load Qh,max to the control system. The control system may then use the display/monitor to indicate the information, such as in method step c), to the boiler operator, such as, by displaying the information.

The EDGE server 203 may be configured to reduce amount of measurement data that is passed to the remote computing system.

A combustion boiler computation system comprises a group of combustion boilers 10, each combustion boiler 10 comprising a boiler control system (CS) comprising an EDGE server (203) system which is configured to process the real-time measurement results for currently monitored process data and/or current load, namely by filtering, averaging, and/or computing trends, and send the processed real-time measurement results to a remote computing system. The remote computing system is preferably a cloud-based computing system, configured to receive data processed from real time measurement results and to compute data using a numerical boiler model for each of the combustion boilers 10, and to return computation results for each of the combustion boilers 10. The boiler control system may be configured to adapt its function based on the computation results. The computing system is preferably configured to find such a numerical value Q _{h, candidate} for a current computational maximum boiler momentary load Q _h,max for which at least one flue gas factor dfi computed using currently monitored process data with a numerical model of the boiler that fulfills an acceptance condition, and selecting the numerical value Q _{h, candidate} as the current computational maximum boiler momentary load Q _h,max .

The boiler computation system may be configured to adapt or calibrate a numerical model for a boiler using processed measurement data for the combustion boiler 10. Alternatively or in addition, the boiler computation system may be configured to adapt or calibrate a numerical model for a combustion boiler 10 using processed measurement data collected also from other combustion boilers 10.

FIG 5 shows a modification of the method shown in FIG 4. Steps LI, L3, L7, L9 are the same as steps K1, K3, K9, K11, respectively, but in step L5, the product gas factors df _i can be directly computed for all heat transfer surfaces 20 _i : if the temperatures T _G,in,i are measured using the respective temperature sensors 21 _i , the back-calculation will not be necessary and thus the step K7 can be omitted in the method illustrated in FIG 5.

FIG 6 shows in step N1 the use of possible inputs to the numerical boiler model. In step N3 the Q _h,max is computed numerically using the boiler model, and in step N5, the estimated maximum load Qh,max is presented to boiler operator via a specific user interface (UI), preferably via display/monitor 202.

FIG 7 shows boiler momentary load Qh and computed current computational maximum boiler momentary load Q _{h, max} , as well as the effect of using the method according to the invention during a test period. During the 10 day test period, the 120 MW _th boiler power obtained in average a 3 to 6 MWth higher load as outside the test period. FIG 8 illustrates the 10 day test period in more detail. In other words, in the control method applied to a boiler, the current computational maximum boiler momentary load Q _h,max of the combustion boiler is estimated using a numerical model using determined boiler's operating parameters. The current boiler load Qh is computed using steam circuit measurement data.

Then, if the boiler load Qh is smaller than the current computational maximum boiler momentary load Q _h,max , it is i) indicated to the boiler operator that the boiler load may be increased, and/or ii) the boiler load is automatically increased. Alternatively or in addition, if the boiler load Qh is larger than the boiler maximum momentary load Q _h,max , it is i) indicated to the boiler operator that the boiler load exceeds the boiler maximum momentary load, and/or ii) the boiler load is automatically reduced.

Above the invention has been explained referring to a combustion process in CFB and BFB. It is clear that the invention is applicable to for example in CFB and BFB gasifiers as well, where instead of flue gas, a cooled product gas is formed for use as desired. Naturally in gasification application the product gas is stored or delivered to a desired further processing instead of admitting to atmosphere through a stack.

The heat releasing reactor may also be for example a waste material gasifier reactor, or a carbon capture reactor

The process requires controllable input flows 22,23 arranged at suitable locations into the reactor 10, at least as follows.

Thermochemical reactor, CaO (generally alkali metal oxides) hyration i. Reactant input, such as CaO, hydration reactor ii. H ₂ O(steam) input iii. Ca(OH) ₂ dehydration reactor iv. Air input

Waste gasifier reactor i. Corresponding input flows as in combustion process, ii. CO and H ₂ and some minor gasification products.

According to another embodiment of the invention, when the reactor is a so-called carbon capture reactor, including a fluidized bed carbonator and calciner connected with each other suitably, which may be arranged in a CFB reactor as shown in the FIG 1. The reactor configured to reduce particularly CO ₂ from a gas to be cleaned (carbonator)by reaction with CaO and producing substantially pure CO ₂ (calciner) by calcination of CaCO ₃ . The gas flow determination may be provided in a same manner as in case of the combustion process, considering that fuel is being burned in a calciner, including gas flow of oxygen and the gas to be cleaned.

FIG 9 depicts a heat releasing reactor 10 according to an embodiment of the invention. The heat releasing reactor may be for example thermochemical reactor, such as CaO (generally alkali metal oxides) hydration reactor. The reactor 10 comprises a reactor chamber 12 that is enclosed by walls 13, which may be optionally cooled walls connected to fluid circuit for extracting heat from process practiced in the reactor chamber - depending on the practical application.

Reactant is fed into the reaction chamber 12 via a reactant inlet 22. The reaction can be adjusted by controlling the reactant feed inlet 22, and generally by controlling the process variables which relate the heat releasing process of the reactor 10.

Reaction products, which may generally be referred to as product gas, flows from the reactor chamber 12 to a gas flow channel 16. The reactor 10 may be also provided with an outlet 22' for solid material which has at least partially reacted in the reactor chamber 12. It is described here as a vertical pass, but it may as well be differently designed, such as horizontally. From the product gas flow channel, the product gas is led to further processing 19, which may comprise simple storage, or a gas delivery pipework.

The product gas flow channel 16 comprises a number of heat transfer surfaces 21i (where 1 = 1, 2, 3, ..., k, where k is the number of heat transfer surfaces). In FIG 9, heat transfer surfaces 21 ₁ , 21 ₂ , 21 ₃ , ..., 21 _k-1 , 21 _k are illustrated. The actual number of different heat transfer surfaces in each of these components, for example, may be selected for each reactor 10 differently according to actual needs.

Product gas exiting the last heat transfer surface 21k will be in exit temperature T _{G, exit} . This temperature is measured with temperature sensor 20 _k .

According to one aspect, the temperatures before and after each heat transfer surface 21i (T _G,in,i, T _G,in,i+i, respectively) can be measured with respective temperature sensors 20i (where I = 1, 2, 3, ..., k-1, k).

According to another aspect, and preferably, these temperatures however do not necessarily need to be measured. It will suffice to know the flue gas exit temperature T _{G, exit} . The temperatures before and after each preceding heat transfer surface 21 _i (T _G,in,i, T _FG,in,i+1 ) can be obtained numerically. This will be explained further below.

The reactor 10 is equipped with a plurality of sensors 40, and computer units, and only some of the sensors and computer units are presented for sake of clarity. For example, there may be sensors 40 for determining flow rate of a reactants into the reactor 10, sensor 40 indicating temperature at one or more locations in the reactor 10, sensors 40 for indicating pressure in the reactor, depending on the actual practical application of the reactor 10.

Process data may be collected from the sensors by distributed control system (DCS) 201. This is similar to that disclosed in connection with the FIG 1, of course the combustion boiler network is instead a heat releasing network, operating in corresponding manner.

It is obvious to the skilled person that, along with the technical progress, the basic idea of the invention can be implemented in many ways. The invention and its embodiments are thus not limited to the examples and samples described above but they may vary within the contents of patent claims and their legal equivalents.

In addition, or instead of using above mentioned specific empirical equations, it is possible to utilize artificial intelligence tools and/or neural network in the numerical model computations.

In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word "comprise" or variations such as "comprises" or "comprising" is used in an inclusive sense, i.e. to specify the presence of the stated feature but not to preclude the presence or addition of further features in various embodiments of the invention.