Trouble emulating microcontroller on Simulink

Question

I am trying to simulate a microcontroller controlled boost converter on Simulink. Because Simulink doesn't have a microcontroller, I wrote a MATLAB function that emulates one.

In case you're wondering why I'm doing this instead of using a PID control block (which is native to Simulink), it's because I'm going to make an STM32 controlled DC-DC converter, which is why I am trying to emulate a microcontroller.

Before I explain the problem, let me explain the MCU function. The MCU function has two inputs: ADC0 and ADC1. These inputs read an analog value and convert it into a 12-bit equivalent value (which you will see in a bit). I am feeding the output voltage to ADC0 and the reference voltage to ADC1.

The MUC function has one output: Dcontrol. This outputs a value between 0 and 1, which is fed to a PWM block. The output Dcontrol is calculated using a praportional-integral (PI) algorithm.

function Dcontrol = MCU(ADC0, ADC1)
    
    %Initialising the integral error term to zero
    err_I = 0; 
    %This, as I shall demonstrate, causes a problem.

    %Converting inputs into 12-bit equivalent value
    Vout = (ADC0/3.3)*4096;
    Vref = (ADC1/3.3)*4096;  

    %Calculating the error, i.e., difference between set-point and
    %measured-value
    err = Vref - Vout;
    
    %Calculating the proporational error
    err_P = err;
    %Calculating the integral error
    err_I = err_I + err;
    
    %Saturating err_I to prevent integral windup
    if(err_I > 800)
        err_I = 800;
    end

    if(err_I < 0)
        err_I = 0;
    end

    %Saturating err_P
    if(err_P > 100)
        err_P = 100;
    end

    if(err_P < -100)
        err_P = -100;
    end

    %Calculating PI
    PI = err_I + err_P;

    %Disallowing a negaive PI value
    if(PI < 0)
        PI = 0;
    end

    %The dutycycle control output is PI/1000, which gives a max dutycycle of 0.9 
    Dcontrol = PI/1000;
end

The equivalent circuit using the native PI control block looks like this:

I'm just showing this for clarity.

Coming back to the microcontroller circuit, I've set the reference voltage to 3V, which should give me 30V at the output. But when I run the simulation, this happens:

My output voltage is around 26V. This is due to the following line in the code

%Initialising the integral error term to zero
err_I = 0; 

Every time the MCU function runs, the integral error gets initialised to zero. This results in the error term (and hence Dcontrol) getting improperly calculated. If I change my input voltage, the output voltage also changes. This is due to the integral term getting reinitialised. If is remove that line, I get an error:

Undefined function or variable 'err_I'. The first assignment to a local variable determines its class.

So my question is - is there a way to initialise the integral error term during runtime only, the way it's possible with the native PID block?

In other words, I want to modify the code to have this effect

    %Initialising the integral error term to zero
    void setup()
    {
    err_I = 0; 
    }

   void loop()
   { 
   %Converting inputs into 12-bit equivalent value
    Vout = (ADC0/3.3)*4096;
    Vref = (ADC1/3.3)*4096;  

    %Calculating the error, i.e., difference between set-point and
    %measured-value
    err = Vref - Vout;
    
    %Calculating the proporational error
    err_P = err;
    %Calculating the integral error
    err_I = err_I + err;
    
    %Saturating err_I to prevent integral windup
    if(err_I > 800)
        err_I = 800;
    end

    if(err_I < 0)
        err_I = 0;
    end

    %Saturating err_P
    if(err_P > 100)
        err_P = 100;
    end

    if(err_P < -100)
        err_P = -100;
    end

    %Calculating PI
    PI = err_I + err_P;

    %Disallowing a negaive PI value
    if(PI < 0)
        PI = 0;
    end

    %The dutycycle control output is PI/1000, which gives a max dutycycle of 0.9 
    Dcontrol = PI/1000;
end
}

Answer 1

Use an S-Function instead of .m code block.

With S-Function you can write C code with one-time-run-only functions, there you can declare variables once to initialize their values and then have them in the main loop function.