10.3.1. Mathematical Representation

Integral Control (I) seeks to eliminate steady-state error by considering the accumulation of past errors. The integral action ensures that any persistent error over time results in a change in the control input, gradually driving the error to zero.

Integral Control is designed to address a specific issue in control systems known as steady-state error. This is the difference between what we want our output to be and what it actually is when the system has reached a steady state. The integral control takes all past errors into account rather than just the current error. By integrating (or summing) these errors over time, the controller can adjust the control input to ensure that any long-term errors are corrected. This is particularly useful when there are small but persistent discrepancies between the desired output and the actual output.

Imagine a person trying to perfectly balance a seesaw. If one side is consistently lower, they add weight to that side to bring it back to the center. As time goes on, they keep track of how tilted the seesaw was for every moment until they adjust it just right. Essentially, just like the person remembers how much effort they put in to balance the seesaw, integral control remembers past errors to maintain the desired performance.

Mathematical Representation: u(t) = K_i ∫_0^t e(τ)dτ where: ● u(t) is the control input. ● K_i is the integral gain. ● e(t) is the error signal. ● The integral term accumulates the error over time, making corrections based on past errors.

The mathematical representation of integral control helps us understand how the control input is calculated. Here, 'u(t)' represents the output of the control system at time 't'. The term 'K_i' is a constant that scales how effectively we respond to the accumulated past errors. The integral from 0 to 't' of 'e(τ)' means we are summing up all the errors from the starting time (0) to the current time 't'. This accumulated sum drives the control input. The more significant the integral gain 'K_i', the more influence past errors have on the control action at any moment.

Think of this formula like a bank account where you deposit every penny you save over time. Every month (or time unit), you keep adding small amounts (the errors) to your total savings. The higher your 'savings rate' (integral gain 'K_i'), the more money you'll accumulate over time, allowing you to make larger investments (control actions) based on that total. Just as a higher saving rate speeds up your ability to reach a financial goal, a higher integral gain allows faster correction of the persistent errors.

Implementation Steps:

1. Measure the error e(t) between the desired setpoint and actual output.
2. Integrate the error over time to accumulate the error (in software, this is typically implemented as a summation of discrete errors over each time step).
3. Apply the integral gain K_i to the accumulated error to compute the control input u(t).
4. Adjust the system (e.g., heating element, actuator) based on the control input.

The steps for implementing integral control provide a clear guide on how to put this control strategy into practice. First, we need to measure the error, which tells us how far we are from our desired output. Next, we integrate this error, meaning we keep a running total over time. This can be done in software by adding up the errors at each time step. Once we have an accumulated error, we multiply it by the integral gain to determine how much adjustment we need to make to the control input. Finally, we use this control input to adjust the system, such as changing the power of a heater or the position of a valve.

Imagine a teacher keeping track of a student's homework errors over the school year. Each time the student misunderstands a topic, the teacher notes it down. At the end of each week, the teacher reviews all the missed points (integrating past errors). The more errors the student accumulates, the more help the teacher provides in the coming lessons (applying integral gain) to ensure the student understands the material by the end of the course (adjusting the teaching based on accumulated errors).

--